CN119672654A - Vehicle information collection and management method and system - Google Patents
Vehicle information collection and management method and system Download PDFInfo
- Publication number
- CN119672654A CN119672654A CN202510181727.2A CN202510181727A CN119672654A CN 119672654 A CN119672654 A CN 119672654A CN 202510181727 A CN202510181727 A CN 202510181727A CN 119672654 A CN119672654 A CN 119672654A
- Authority
- CN
- China
- Prior art keywords
- image
- vehicle
- license plate
- contour
- shielding
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Image Analysis (AREA)
Abstract
The invention provides a vehicle information recording management method and system, and relates to the technical field of vehicle management, wherein the method comprises the steps of acquiring an acquisition image of a monitored vehicle; the method comprises the steps of acquiring an image, obtaining the vehicle characteristics of a monitored vehicle according to the acquired image, obtaining the license plate outline of the monitored vehicle according to the vehicle characteristics through an edge detection algorithm, processing the license plate outline through morphological operation according to the license plate outline to obtain an enhanced image of the license plate outline, dividing the enhanced image of the license plate outline into a non-shielding area and a shielding area through a deep learning model according to the enhanced image of the license plate outline, repairing according to the non-shielding area through context information to obtain the shielding image of the shielding area, obtaining the license plate number of the monitored vehicle according to the image and the shielding image of the non-shielding area, and recording the license plate number and the vehicle type of the monitored vehicle. The invention realizes the rapid and accurate recording of the vehicle information by combining the advanced image processing technology and the deep learning algorithm.
Description
Technical Field
The invention relates to the technical field of vehicle management, in particular to a vehicle information recording management method and system.
Background
The vehicle information plays an important role in monitoring and managing the vehicle, and the acquisition of the vehicle information, such as a vehicle disc number, a vehicle type and the like, is helpful for a vehicle management department to better manage the vehicle so as to record and manage the technical condition, insurance information, drivers and the like of the vehicle, thereby improving traffic safety and preventing traffic accidents.
At present, the traditional vehicle information recording method mainly depends on manual identification and recording, but the manual operation efficiency is relatively low, and particularly in the area with large traffic flow, the information of each vehicle is difficult to record quickly and accurately. In addition, when the vehicle license plate is deliberately blocked or the images are unclear due to the problems of dirt, insufficient light, camera angles and the like, the difficulty of manual identification is further increased, and the error rate is also improved.
Disclosure of Invention
The invention solves the problem of how to improve the efficiency and the accuracy of the vehicle information recording.
In order to solve the above problems, the present invention provides a vehicle information recording management method, including:
Acquiring an acquisition image of a monitoring vehicle;
Obtaining vehicle characteristics of the monitored vehicle according to the acquired image, wherein the vehicle characteristics comprise vehicle types, aspect ratio and color distribution of the monitored vehicle;
Obtaining the license plate outline of the monitoring vehicle through an edge detection algorithm according to the aspect ratio and the color distribution;
Processing the license plate contour through morphological operation according to the license plate contour to obtain an enhanced image of the license plate contour;
Dividing the enhanced image of the license plate contour into a non-shielding region and a shielding region through a deep learning model according to the enhanced image of the license plate contour;
Repairing according to the non-shielding region by context information to obtain a shielding image of the shielding region;
Obtaining a license plate number of the monitoring vehicle according to the image of the non-shielding area and the shielding image;
and recording the license plate number and the vehicle type of the monitored vehicle.
Optionally, the obtaining the vehicle feature of the monitored vehicle according to the acquired image includes:
Preprocessing the acquired image to obtain the processed acquired image;
extracting features of the processed acquired images through an image processing technology to obtain the contour of the monitored vehicle, the contour of the vehicle window, the contour of the vehicle lamp and the contour of the vehicle wheel;
obtaining vehicle characteristics according to the contour of the monitoring vehicle, the contour of the vehicle window, the contour of the vehicle lamp and the contour of the vehicle wheel through a graph rolling network;
And classifying the vehicle characteristics according to the vehicle management data to obtain the vehicle type corresponding to the monitored vehicle.
Optionally, the obtaining, according to the aspect ratio and the color distribution, the license plate profile of the monitored vehicle through an edge detection algorithm includes:
performing binarization processing on the acquired image to obtain a binarized image of the acquired image;
obtaining a plurality of geometric features of the monitoring vehicle through findcontours functions according to the binarized image;
obtaining an aspect ratio of each of the geometric features;
and screening the geometric features according to the aspect ratio, the aspect ratio of the geometric features and the color distribution to obtain the license plate outline.
Optionally, the screening the geometric feature according to the aspect ratio, the aspect ratio of the geometric feature and the color distribution to obtain the license plate outline includes:
Obtaining the corresponding color of the geometric feature in the acquired image according to the color distribution;
screening the geometric features according to the aspect ratio, the preset license plate aspect ratio and the preset color;
and judging the outline of the geometric feature as the license plate outline when the difference between the aspect ratio of the geometric feature and the aspect ratio of the preset license plate is smaller than a preset threshold value and the color difference between the color and the preset color is smaller than a preset color difference threshold value.
Optionally, the processing the license plate contour through morphological operation according to the license plate contour to obtain an enhanced image of the license plate contour includes:
dividing the area where the license plate outline is located in the binarized image to obtain a license plate binarized image;
Performing expansion operation on the license plate binarization image to obtain an expanded binarization image;
and performing corrosion operation on the expanded binary image to obtain a corroded binary image, and taking the corroded binary image as the enhanced image of the license plate outline.
Optionally, the dividing the enhanced image of the license plate contour into a non-occlusion area and an occlusion area by a deep learning model according to the enhanced image of the license plate contour includes:
inputting the enhanced image into the deep learning model, and obtaining local characteristics of the enhanced image through a convolution layer of the deep learning model;
inputting the local features into a pooling layer of the deep learning model, reducing the space dimension of the local features, and obtaining the feature representation of the enhanced image through an activation function layer;
optimizing the characteristic representation through an cross-ratio loss function to obtain the optimized characteristic representation;
And carrying out pixel classification on the enhanced image according to the optimized characteristic representation to obtain the non-shielding region and the shielding region.
Optionally, acquiring context information of the non-occlusion region, wherein the context information comprises image information of a surrounding preset region of the non-occlusion region;
Obtaining the overall structural characteristics of the license plate outline and the characteristics of the non-shielding region according to the image information of the surrounding preset region of the non-shielding region;
inputting the integral structure characteristics and the characteristics of the non-shielding area into a self-encoder to generate an estimated image of the shielding area;
splicing the image of the non-shielding area with the estimated image, and judging whether the spliced image is available or not;
if yes, the estimated image is used as the shielding image of the shielding area.
Optionally, the obtaining the license plate number of the monitored vehicle according to the image of the non-occlusion area and the occlusion image includes:
splicing the image of the non-shielding area and the shielding image to obtain a clear image corresponding to the license plate contour;
performing character segmentation according to the clear image to obtain the geometric characteristics of each character in the clear image;
And identifying each character according to the geometric features to obtain the license plate number of the monitoring vehicle.
Optionally, the recording the license plate number of the monitored vehicle and the vehicle type includes:
correlating the license plate number with the vehicle type to obtain a comprehensive data record of the monitored vehicle;
automatically generating a text file corresponding to the comprehensive data record through an automatic script, and recording the text file into a database.
The invention also provides a vehicle information recording management system, which comprises:
The acquisition unit is used for acquiring an acquisition image of the monitoring vehicle;
The feature extraction unit is used for obtaining the vehicle features of the monitoring vehicle according to the acquired image, wherein the vehicle features comprise the vehicle type, the aspect ratio and the color distribution of the monitoring vehicle;
The contour extraction unit is used for obtaining the license plate contour of the monitoring vehicle through an edge detection algorithm according to the aspect ratio and the color distribution;
The image processing unit is used for processing the license plate contour through morphological operation according to the license plate contour to obtain an enhanced image of the license plate contour;
Dividing the enhanced image of the license plate contour into a non-shielding region and a shielding region through a deep learning model according to the enhanced image of the license plate contour;
The image restoration unit is used for restoring according to the non-shielding area through the context information to obtain a shielding image of the shielding area;
the identification unit is used for obtaining the license plate number of the monitoring vehicle according to the image of the non-shielding area and the shielding image;
and the data importing unit is used for inputting the license plate number and the vehicle type of the monitored vehicle.
According to the vehicle information recording management method and system, the system can rapidly capture the real-time state of the vehicle by collecting and monitoring the image of the vehicle. Then, vehicle characteristics including vehicle type, aspect ratio and color distribution are extracted by using an image processing technology, and the characteristics provide basic information for subsequent license plate recognition. And then through an edge detection algorithm, the system can accurately identify the outline of the license plate, morphological operation further enhances the definition of the outline of the license plate, and provides high-quality input for the identification of a deep learning model. The application of the deep learning model is a key for improving efficiency and accuracy, and can distinguish a non-shielding area and a shielding area in a license plate image, which is particularly important in practical application, because the license plate can be shielded by other objects, and the recognition accuracy is affected. The non-shielding area is repaired by utilizing the context information to carry out an image repair technology, and the shielded license plate part can be effectively restored, so that the integrity and the accuracy of license plate recognition are improved. Finally, by combining the non-shielding area and the repaired shielding area, the system can accurately acquire the license plate number and input the license plate number and the vehicle type information into a database. The system can adapt to different environments and illumination conditions, and can keep high recognition accuracy even under complex or unfavorable conditions. Not only reduces manual intervention and error rate, but also improves the speed of information processing, so that the recording and management of vehicle information are more efficient and accurate. By integrating and inputting information such as license plate numbers, vehicle types and the like, the system provides a comprehensive data base for subsequent data analysis and decision making, realizes quick response of updating of vehicle information, and provides instant data support for traffic management and vehicle scheduling. The invention combines advanced image processing technology and deep learning algorithm to realize rapid and accurate recording of vehicle information, provides powerful technical support for traffic management and vehicle monitoring, and improves the efficiency and accuracy of vehicle information recording.
Drawings
FIG. 1 is a flowchart of a method for managing recording of vehicle information according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a recording management system based on vehicle information according to an embodiment of the present invention.
Detailed Description
In order that the above objects, features and advantages of the invention will be readily understood, a more particular description of the invention will be rendered by reference to specific embodiments thereof which are illustrated in the appended drawings.
Referring to fig. 1, the present invention provides a vehicle information recording management method, including:
An acquisition image of a monitored vehicle is acquired.
Specifically, the first step of vehicle information recording is image acquisition, which is usually completed by a high-definition camera, and is generally deployed at key monitoring points, such as traffic intersections, parking lot entrances or toll booths, and the like. The cameras can capture images of the passing vehicles in real time, and original data is provided for subsequent vehicle feature analysis. The quality of the image directly affects the accuracy of subsequent processing, and therefore, the acquired image needs to have sufficient resolution and sharpness.
And obtaining the vehicle characteristics of the monitored vehicle according to the acquired image, wherein the vehicle characteristics comprise the vehicle type, the aspect ratio and the color distribution of the monitored vehicle.
Specifically, after the vehicle image is acquired, a step of extracting the vehicle features follows. This step involves image processing and computer vision techniques, which analyze the images through algorithms to identify specific attributes of the vehicle, such as vehicle type, aspect ratio, and color distribution. These features are critical to the identification and classification of vehicles. For example, a model may be trained using a machine learning algorithm to identify features of different vehicle models, or a color clustering algorithm may be used to analyze the color distribution of the vehicle.
And obtaining the license plate outline of the monitoring vehicle through an edge detection algorithm according to the aspect ratio and the color distribution.
In particular, edge detection is a key technology in image processing for identifying edge information in an image. In vehicle information recording, edge detection algorithms are used to identify the outline of a license plate. Common edge detection algorithms include Sobel, prewitt, canny, etc. These algorithms determine the location of the edge by calculating the gradient change of the pixels in the image. For example, the Canny algorithm is widely used in license plate detection with its high detection accuracy and robustness to noise.
And processing the license plate contour through morphological operation according to the license plate contour to obtain an enhanced image of the license plate contour.
In particular, morphological operations are another important class of techniques in image processing, which are analyzed and processed based on the shape and structure of the image. In the processing of license plate contours, morphological operations such as swelling and corrosion are used to enhance the recognizability of the license plate. The expansion operation can fill in small gaps between license plate characters, and the corrosion operation can remove noise and unnecessary details on the license plate. These operations help to improve accuracy of license plate recognition.
And dividing the enhanced image of the license plate contour into a non-shielding area and a shielding area through a deep learning model according to the enhanced image of the license plate contour.
In particular, deep learning techniques have revolutionized the field of image recognition. In vehicle information recording, a deep learning model is used to distinguish between non-occluded and occluded areas in a license plate image. This typically involves the use of a trained convolutional neural network to identify and classify different regions in the image. For example, a deep learning model may be trained to recognize letters and numbers on a license plate while ignoring incomplete portions of the image due to occlusion or other reasons.
And repairing according to the non-shielding area by the context information to obtain a shielding image of the shielding area.
In particular, image restoration techniques by contextual information aim to recover damaged or incomplete portions of an image. In license plate recognition, if a portion of the license plate is occluded, the image restoration technique may predict and fill in the missing portion based on surrounding context information. The method can improve the integrity and accuracy of license plate recognition. For example, if a character of the license plate is partially obscured, surrounding visible characters and known license plate formats may be used to infer the missing portion.
And obtaining the license plate number of the monitoring vehicle according to the image of the non-shielding area and the shielding image.
In particular, in a preferred embodiment of the present invention, the identification of license plate numbers generally relies on optical character recognition techniques. Optical character recognition techniques are capable of converting characters in an image into an editable text format. In license plate recognition systems, optical character recognition techniques are used to extract license plate numbers from processed license plate images. This typically involves using specific optical character recognition techniques to recognize characters in the image and convert them to a standard license plate number format.
And recording the license plate number and the vehicle type of the monitored vehicle.
Specifically, the last step is to input the identified license plate number and the vehicle type information into a database. This step involves database management techniques, including storage, retrieval and management of data. The vehicle information is input not only with accuracy, but also with an efficient data organization mode, so that subsequent inquiry and analysis are facilitated. For example, vehicle information may be stored in categories by time, location, or vehicle type to facilitate statistical analysis or traffic management decisions.
According to the vehicle information recording management method, the system can rapidly capture the real-time state of the vehicle by collecting and monitoring the image of the vehicle. Then, vehicle characteristics including vehicle type, aspect ratio and color distribution are extracted by using an image processing technology, and the characteristics provide basic information for subsequent license plate recognition. And then through an edge detection algorithm, the system can accurately identify the outline of the license plate, morphological operation further enhances the definition of the outline of the license plate, and provides high-quality input for the identification of a deep learning model. The application of the deep learning model is a key for improving efficiency and accuracy, and can distinguish a non-shielding area and a shielding area in a license plate image, which is particularly important in practical application, because the license plate can be shielded by other objects, and the recognition accuracy is affected. The non-shielding area is repaired by utilizing the context information to carry out an image repair technology, and the shielded license plate part can be effectively restored, so that the integrity and the accuracy of license plate recognition are improved. Finally, by combining the non-shielding area and the repaired shielding area, the system can accurately acquire the license plate number and input the license plate number and the vehicle type information into a database. The system can adapt to different environments and illumination conditions, and can keep high recognition accuracy even under complex or unfavorable conditions. Not only reduces manual intervention and error rate, but also improves the speed of information processing, so that the recording and management of vehicle information are more efficient and accurate. By integrating and inputting information such as license plate numbers, vehicle types and the like, the system provides a comprehensive data base for subsequent data analysis and decision making, realizes quick response of updating of vehicle information, and provides instant data support for traffic management and vehicle scheduling. The invention combines advanced image processing technology and deep learning algorithm to realize rapid and accurate recording of vehicle information, provides powerful technical support for traffic management and vehicle monitoring, and improves the efficiency and accuracy of vehicle information recording.
Optionally, the obtaining the vehicle feature of the monitored vehicle according to the acquired image includes:
Preprocessing the acquired image to obtain the processed acquired image;
extracting features of the processed acquired images through an image processing technology to obtain the contour of the monitored vehicle, the contour of the vehicle window, the contour of the vehicle lamp and the contour of the vehicle wheel;
obtaining vehicle characteristics according to the contour of the monitoring vehicle, the contour of the vehicle window, the contour of the vehicle lamp and the contour of the vehicle wheel through a graph rolling network;
And classifying the vehicle characteristics according to the vehicle management data to obtain the vehicle type corresponding to the monitored vehicle.
Specifically, first, the extraction process of the vehicle features starts with preprocessing of the captured image. This step is critical because it can improve the accuracy and efficiency of subsequent processing. Preprocessing typically includes denoising, image enhancement, histogram equalization, etc., in order to remove noise from the image and enhance the recognizability of the vehicle features. For example, salt and pepper noise can be removed using a median filter, while histogram equalization can improve the contrast of the image, making the contours, windows, lights, wheels, etc. of the vehicle more prominent. Next, feature extraction is performed on the preprocessed image by an image processing technique. This step may employ an edge detection algorithm, such as the Canny operator, to identify the contours of the vehicle, a Hough transform to detect rounded parts of the vehicle, such as the vehicle lights and wheels, and a color segmentation technique to identify the regions of the vehicle window. The key point of feature extraction is to be able to accurately identify and locate the various components of the vehicle, providing reliable data for subsequent feature classification. Finally, the application of the graph rolling network is based on these extracted features. The graph convolutional network is a deep learning model capable of processing graph structure data and is well suited for processing interrelated features such as different parts of a vehicle. The graph convolutional network can learn a richer and more robust representation of vehicle features by aggregating the feature information of each vehicle part and combining the topological relationships in the graph structure. For example, the graph roll-up network may effectively integrate the features of the contours, windows, lights, and wheels of the vehicle to achieve accurate classification of the vehicle. In a preferred embodiment of the invention, the acquired vehicle image is first preprocessed, using wavelet transform to remove noise and preserve edge information. Then, edge detection is carried out through a Canny operator to obtain the outline of the vehicle, the vehicle lamp and the vehicle wheel are identified through Hough transformation, and the vehicle window area is identified through color space conversion and cluster analysis. These features are then input into a trained, graph roll-up network model that has been learned over a large number of vehicle images to be able to identify and classify different types of vehicle models.
In the embodiment, the efficiency of vehicle information processing is greatly improved through automatic feature extraction and classification, and the requirement of manual intervention is reduced. And secondly, the accuracy of feature extraction is improved by image preprocessing and advanced image processing technology, so that the accuracy of vehicle type recognition is improved. Reduces manual intervention, reduces errors caused by human factors, and improves processing speed and efficiency. Through continuous learning and optimization of the deep learning model, the system can adapt to various complex environments and conditions, and high-accuracy recognition capability is maintained. The invention not only improves the intelligent level of vehicle information management, but also provides powerful technical support for the fields of traffic monitoring, vehicle scheduling, safety supervision and the like.
Optionally, the obtaining, according to the aspect ratio and the color distribution, the license plate profile of the monitored vehicle through an edge detection algorithm includes:
performing binarization processing on the acquired image to obtain a binarized image of the acquired image;
obtaining a plurality of geometric features of the monitoring vehicle through findcontours functions according to the binarized image;
obtaining an aspect ratio of each of the geometric features;
and screening the geometric features according to the aspect ratio, the aspect ratio of the geometric features and the color distribution to obtain the license plate outline.
Specifically, the acquired image is first subjected to binarization processing, which is to convert the image into an image having only two pixel values of black and white, wherein white represents a foreground (an area of interest such as a vehicle or a license plate) and black represents a background. Binarization may be achieved by setting a threshold above which pixels are considered white and below which pixels are considered black. This step is crucial for subsequent edge detection and feature extraction, as it simplifies the image, making it easier for the algorithm to identify the region of interest. And secondly, performing contour detection on the binarized image by using findcontours functions. This function can identify all white areas in the image and extract their boundaries (contours). Each contour may be represented by a series of points that define the outline of the object. In this process, a plurality of geometric features, such as the area, perimeter, center point, etc., of each contour may be acquired. Finally, by analyzing these geometric features, in particular their aspect ratio, the regions that are likely to be license plates can be screened. Aspect ratio is an important feature because the aspect ratio of license plates is typically within a certain range, and the detected contours are screened in combination with color distribution information. For example, a reasonable aspect ratio range may be set, and only contours that match this range are considered potential license plates to further confirm the location of the license plates. For example, if the aspect ratio of a geometric feature matches the standard scale of a license plate and the color distribution matches the typical color of a license plate, then the feature is likely to represent a license plate. In a preferred embodiment of the present invention, the acquired vehicle image is first binarized, and the Otsu method may be used to automatically determine the threshold value to adapt to the image under different lighting conditions. Then, findcontours functions in the OpenCV library are applied to extract the contours of all white areas. Contours with aspect ratios between 1.5 and 2.5 are screened out by calculating the aspect ratio of each contour, which generally corresponds to the aspect ratio of the license plate. In addition, color histogram analysis may be used to further screen out contours that do not match license plate color for those color distributions.
In the embodiment, the image information is greatly simplified through screening of the aspect ratio and analysis of the color distribution and binarization processing of the image, the complex image is converted into a black-and-white image which is easier to analyze through thresholding operation, all connected white areas in the image can be accurately positioned, rich geometric information is provided for extracting license plates, the accuracy of license plate detection is further improved, and the possibility of false detection is reduced.
Optionally, the screening the geometric feature according to the aspect ratio, the aspect ratio of the geometric feature and the color distribution to obtain the license plate outline includes:
Obtaining the corresponding color of the geometric feature in the acquired image according to the color distribution;
screening the geometric features according to the aspect ratio, the preset license plate aspect ratio and the preset color;
and judging the outline of the geometric feature as the license plate outline when the difference between the aspect ratio of the geometric feature and the aspect ratio of the preset license plate is smaller than a preset threshold value and the color difference between the color and the preset color is smaller than a preset color difference threshold value.
Specifically, first, the system will analyze the color of each geometric feature in the captured image. This may be done by converting each geometric feature in the image to an appropriate color space (e.g., RGB) to more accurately represent the colors and make the comparison. The system then screens the geometric features according to the predetermined aspect ratio and color of the license plate. The predetermined aspect ratio is based on the standardized dimensions of the license plate, and the predetermined color is a typical color of the license plate, such as white or yellow. During the screening process, the system compares the aspect ratio of each geometric feature to a predetermined license plate aspect ratio. If the difference is less than a preset threshold, this indicates that the geometric feature may be a license plate. Meanwhile, the system also calculates the chromatic aberration between the geometric characteristic color and the preset color. The calculation of the color difference may be achieved by a number of methods, for example using euclidean distance to measure the proximity of two colors in the color space. This further increases the likelihood that the geometric feature is a license plate if the color difference is also less than a preset color difference threshold. In a preferred embodiment of the present invention, assuming a preset license plate aspect ratio of 3:1, the preset color is white and the color difference threshold is set to 30 (in HSV color space). The system first processes the acquired image to extract all possible geometric features. The system then calculates the aspect ratio of each geometric feature and compares it to a preset aspect ratio. For geometric features with aspect ratios between 2.7 and 3.3, the system further analyzes their color. If the color of the geometric feature is within 30 units of the white color range, the system marks it as a potential license plate. In this way, the system can effectively identify the license plate from the image.
In this embodiment, by considering the two dimensions of the aspect ratio and the color, the system can more accurately screen out geometric features conforming to license plate features, so as to reduce interference of other non-license plate objects, and meanwhile, the preset aspect ratio and color threshold can be adjusted according to standards of license plates in different areas or different types, so that the system can adapt to various license plate specifications and colors. The two dimensions of the shape and the color of the geometric feature are comprehensively considered, the possibility of false recognition is reduced, and the accuracy of license plate recognition is improved.
Optionally, the processing the license plate contour through morphological operation according to the license plate contour to obtain an enhanced image of the license plate contour includes:
dividing the area where the license plate outline is located in the binarized image to obtain a license plate binarized image;
Performing expansion operation on the license plate binarization image to obtain an expanded binarization image;
and performing corrosion operation on the expanded binary image to obtain a corroded binary image, and taking the corroded binary image as the enhanced image of the license plate outline.
In particular, in license plate recognition, morphological operations are used to enhance the outline of the license plate, improving the accuracy of recognition. Firstly, performing binarization processing on the acquired image to obtain a binarization contour of the license plate. Then, the white pixels of the license plate area are increased by using expansion operation, so that small cracks or broken strokes possibly existing in the license plate can be filled, and the license plate area is more complete. The dilation operation achieves the "dilation" effect of the image by using a predefined structural element to slide through the image and increment the pixel value by a certain rule (typically assigning the value of the foreground pixel to the background pixel covered by the structural element). Next, the inflated image is subjected to a corrosion operation. The corrosion operation, as opposed to dilation, refines the outline of the license plate by reducing the white pixels at the edges of the image. The erosion operation is performed by using the same structural element, but this time assigning the value of the background pixel to the foreground pixel covered by the structural element, if the foreground pixel under the structural element is not entirely white, the pixel value at that location will be eroded to the background value. This helps to remove noise and burrs from the edges of the license plate, making the outline of the license plate clearer and more accurate. In the preferred embodiment of the invention, the acquired vehicle image is firstly subjected to binarization processing to obtain a clear license plate contour. Then, a 5×5 rectangular structural element is defined, and a dilation operation is performed on the binarized image. After the expansion operation, the same structural elements are used again to perform the etching operation on the image. Through the two continuous morphological operations, a license plate image with a clearer outline and a more complete interior can be obtained.
In the embodiment, the expansion operation enhances the internal integrity of the license plate, fills up possible cracks, fills up possible small cracks or broken strokes in the license plate by adding white pixels in the license plate area, ensures that the license plate area is more complete, and reduces recognition errors caused by image noise or quality problems. Noise at the edge of the license plate is removed through corrosion operation, so that the outline of the license plate is clearer, and the possibility of false recognition is reduced.
Optionally, the dividing the enhanced image of the license plate contour into a non-occlusion area and an occlusion area by a deep learning model according to the enhanced image of the license plate contour includes:
inputting the enhanced image into the deep learning model, and obtaining local characteristics of the enhanced image through a convolution layer of the deep learning model;
inputting the local features into a pooling layer of the deep learning model, reducing the space dimension of the local features, and obtaining the feature representation of the enhanced image through an activation function layer;
optimizing the characteristic representation through an cross-ratio loss function to obtain the optimized characteristic representation;
And carrying out pixel classification on the enhanced image according to the optimized characteristic representation to obtain the non-shielding region and the shielding region.
Specifically, in the license plate recognition system, the enhanced license plate image is first input into the deep learning model. The convolution layer of the model is responsible for extracting the local features of the image. These convolution layers identify patterns in the image, such as edges, textures, etc., through a series of learnable filters. For example, a first layer of convolution may focus on identifying simple edge features, while a deeper layer of convolution may identify more complex shape and structural features. These local features are then fed into the pooling layer, typically using a max-pooling or average pooling technique, to reduce the spatial dimensions of the features while increasing invariance to image displacement. This step reduces the spatial size of the data, thereby reducing the number of parameters and the computational complexity. Next, by activating the function layer, nonlinearities are introduced so that the model can learn and model more complex function mappings. Finally, the feature representation of the model is optimized using the cross-ratio loss function. IoU is an index that measures the degree of overlap of the predicted and real regions, and by minimizing IoU loss between the predicted and real frames, the position and shape of the license plate can be more accurately located. After the optimization is completed, the model can classify the enhanced image at the pixel level according to the characteristic representation, and distinguish a non-shielding region and a shielding region.
In the embodiment, through the hierarchical structure of the deep learning model, the features can be extracted gradually from simple to complex, so that the generalization capability of the model to different types of license plates is improved, and the model can be better adapted to license plate recognition tasks under different environments and conditions. The use of a convolution layer in combination with a pooling layer significantly improves the ability of feature extraction so that the model can capture a more rich and abstract representation of features. The model is optimized by using the cross-correlation loss function, so that the high consistency of the prediction area output by the model and the real license plate area is ensured, the recognition precision of the shielding area is improved, the recognition error caused by shielding is reduced, and the accuracy and the robustness of license plate recognition are improved.
Optionally, acquiring context information of the non-occlusion region, wherein the context information comprises image information of a surrounding preset region of the non-occlusion region;
Obtaining the overall structural characteristics of the license plate outline and the characteristics of the non-shielding region according to the image information of the surrounding preset region of the non-shielding region;
inputting the integral structure characteristics and the characteristics of the non-shielding area into a self-encoder to generate an estimated image of the shielding area;
splicing the image of the non-shielding area with the estimated image, and judging whether the spliced image is available or not;
if yes, the estimated image is used as the shielding image of the shielding area.
Specifically, first, the system analyzes the non-occlusion region and extracts the overall structural features of the license plate, such as edges, colors, textures, etc. of the license plate. These features are then used to infer the possible appearance and properties of the occluded area. And extracting the integral structural characteristics of the license plate outline and the characteristics of the non-shielding region according to the image information of the non-shielding region and the surrounding preset region. The self-encoder is a powerful neural network capable of learning an efficient representation of data, and is particularly suitable for image restoration tasks. In the repair of a license plate occlusion region, the features of the non-occlusion region are compressed from the encoder portion of the encoder into a low dimensional representation, and the decoder portion expands this representation back to the dimensions of the original image, generating an estimated image of the occlusion region. For example, a self-encoder may be designed that uses convolutional neural network layers for both the encoder and decoder to preserve the spatial hierarchy of the image. In practice, a pre-trained self-encoder model may be employed that has been trained on a large number of non-occluded license plate images to learn typical features of a license plate. When the shielded license plate image needs to be repaired, the non-shielded area is input into the self-encoder, and an estimated image of the shielded area is generated. Then, the estimated image is spliced with the image of the non-occlusion area to form a complete license plate image. An estimated image is considered to be available if the generated image visually coincides with the non-occluded region and conforms to typical structural features of a license plate.
In the embodiment, the context information and the self-encoder are used for repairing the shielding area, so that the robustness of the license plate recognition system is improved, the integrity of the license plate is recovered, the license plate can be accurately recognized even under the condition of partial shielding, more complete and reliable visual information is provided for subsequent license plate recognition, and the accuracy of the license plate recognition system is improved due to the consistency of the structure and the characteristics of the repaired image, particularly under the condition that the shielding area is large or the shielding characteristics are not obvious.
Optionally, the obtaining the license plate number of the monitored vehicle according to the image of the non-occlusion area and the occlusion image includes:
splicing the image of the non-shielding area and the shielding image to obtain a clear image corresponding to the license plate contour;
performing character segmentation according to the clear image to obtain the geometric characteristics of each character in the clear image;
And identifying each character according to the geometric features to obtain the license plate number of the monitoring vehicle.
Specifically, stitching the non-occlusion region and the occlusion image is a key step in the license plate recognition process. First, the occlusion region estimated image generated from the encoder is accurately aligned and fused on the edge with the actual non-occlusion region image. This step is needed to ensure that the stitched image is visually consistent, without obvious seams or unnatural changes, to facilitate subsequent character segmentation and recognition. The stitching technique may utilize a seamless stitching algorithm in image processing, such as a feature point matching based stitching method, or use an image fusion technique, such as poisson fusion, to smooth the transition region. And then character segmentation is carried out on the spliced clear images. This step typically involves binarization of the image to highlight the contrast of the characters with the background, and then applying connected domain analysis or edge detection algorithms to identify and separate each character. For example, the Otsu method may be used to automatically binarize and then locate the character by looking up the outline. Geometric features of each character, such as size, shape and location, will be extracted in preparation for character recognition. In the character recognition stage, a deep learning model, such as a convolutional neural network, may be used to recognize each character in the image. Convolutional neural network models have been pre-trained on a large number of license plate character data to be able to recognize different character shapes and styles. The extracted geometric features are input into a convolutional neural network and the model will output the predicted class for each character.
In the embodiment, the integrity of the license plate image is ensured by accurately splicing the non-shielding area and the shielding image, and high-quality input is provided for character segmentation and recognition. Secondly, the accuracy of character segmentation directly influences the final recognition result, and the segmentation accuracy can be improved and the risk of false recognition can be reduced by using an advanced image processing technology.
Optionally, the recording the license plate number of the monitored vehicle and the vehicle type includes:
correlating the license plate number with the vehicle type to obtain a comprehensive data record of the monitored vehicle;
automatically generating a text file corresponding to the comprehensive data record through an automatic script, and recording the text file into a database.
Specifically, a data structure is first required to store license plate number and model information of each monitored vehicle. This data structure is typically a record containing a plurality of fields, for example, in a relational database, a table containing license plate number, model, time stamp, etc. Wherein, the writing of the automation script is an important means for realizing the automatic input of data. This script may be implemented using a variety of programming languages. The core function of the script includes receiving license plate number and vehicle model information as inputs, and then formatting the information into a specific text file format, such as CSV or JSON. In a further preferred embodiment of the present invention, the Python script may use a built-in CSV module to generate CSV files, each row in the file representing a comprehensive data record of a vehicle.
In the embodiment, the requirement of manual input is reduced through an automatic process, and the efficiency and the speed of data input are obviously improved. Meanwhile, the possibility of human input errors is reduced, and the accuracy and consistency of data are ensured. Text files generated using scripts (e.g., CSV or JSON formats) are easy to integrate with other systems or applications, providing good flexibility and extensibility. Not only improves the speed and quality of data processing, but also provides data support for intelligent traffic and vehicle management.
As shown in fig. 2, the present invention further provides a vehicle information recording management system, including:
The acquisition unit is used for acquiring an acquisition image of the monitoring vehicle;
The feature extraction unit is used for obtaining the vehicle features of the monitoring vehicle according to the acquired image, wherein the vehicle features comprise the vehicle type, the aspect ratio and the color distribution of the monitoring vehicle;
The contour extraction unit is used for obtaining the license plate contour of the monitoring vehicle through an edge detection algorithm according to the aspect ratio and the color distribution;
The image processing unit is used for processing the license plate contour through morphological operation according to the license plate contour to obtain an enhanced image of the license plate contour;
Dividing the enhanced image of the license plate contour into a non-shielding region and a shielding region through a deep learning model according to the enhanced image of the license plate contour;
The image restoration unit is used for restoring according to the non-shielding area through the context information to obtain a shielding image of the shielding area;
the identification unit is used for obtaining the license plate number of the monitoring vehicle according to the image of the non-shielding area and the shielding image;
and the data importing unit is used for inputting the license plate number and the vehicle type of the monitored vehicle.
According to the vehicle information recording management system, the real-time state of the vehicle can be rapidly captured by collecting and monitoring the image of the vehicle. Then, vehicle characteristics including vehicle type, aspect ratio and color distribution are extracted by using an image processing technology, and the characteristics provide basic information for subsequent license plate recognition. Through an edge detection algorithm, the system can accurately identify the outline of the license plate, morphological operation further enhances the definition of the outline of the license plate, and provides high-quality input for the identification of a deep learning model. The application of the deep learning model is a key for improving efficiency and accuracy, and can distinguish a non-shielding area from a shielding area in a license plate image, which is particularly important in practical application, because the license plate can be shielded by other objects, and the recognition accuracy is affected. And the non-shielding area is repaired by utilizing the context information, so that the shielded license plate part can be effectively restored, and the integrity and accuracy of license plate recognition are improved. Finally, by combining the non-shielding area and the repaired shielding area, the system can accurately acquire the license plate number and input the license plate number and the vehicle type information into a database. Not only reduces manual intervention and error rate, but also improves the speed of information processing, so that the recording and management of vehicle information are more efficient and accurate. The invention combines advanced image processing technology and deep learning algorithm to realize rapid and accurate recording of vehicle information, provides powerful technical support for traffic management and vehicle monitoring, and improves the efficiency and accuracy of vehicle information recording.
Those skilled in the art will appreciate that all or part of the processes in the methods of the above embodiments may be implemented by a computer program for instructing relevant hardware, where the program may be stored in a non-volatile computer readable storage medium, and where the program, when executed, may include processes in the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (SYNCHLINK) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
It should be noted that in this document, relational terms such as "first" and "second" and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises an element.
The foregoing is only a specific embodiment of the invention to enable those skilled in the art to understand or practice the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A vehicle information recording management method, comprising:
Acquiring an acquisition image of a monitoring vehicle;
Obtaining vehicle characteristics of the monitored vehicle according to the acquired image, wherein the vehicle characteristics comprise vehicle types, aspect ratio and color distribution of the monitored vehicle;
Obtaining the license plate outline of the monitoring vehicle through an edge detection algorithm according to the aspect ratio and the color distribution;
Processing the license plate contour through morphological operation according to the license plate contour to obtain an enhanced image of the license plate contour;
Dividing the enhanced image of the license plate contour into a non-shielding region and a shielding region through a deep learning model according to the enhanced image of the license plate contour;
Repairing according to the non-shielding region by context information to obtain a shielding image of the shielding region;
Obtaining a license plate number of the monitoring vehicle according to the image of the non-shielding area and the shielding image;
and recording the license plate number and the vehicle type of the monitored vehicle.
2. The vehicle information recording management method according to claim 1, wherein the obtaining the vehicle characteristic of the monitoring vehicle from the collected image includes:
Preprocessing the acquired image to obtain the processed acquired image;
extracting features of the processed acquired images through an image processing technology to obtain the contour of the monitored vehicle, the contour of the vehicle window, the contour of the vehicle lamp and the contour of the vehicle wheel;
obtaining vehicle characteristics according to the contour of the monitoring vehicle, the contour of the vehicle window, the contour of the vehicle lamp and the contour of the vehicle wheel through a graph rolling network;
And classifying the vehicle characteristics according to the vehicle management data to obtain the vehicle type corresponding to the monitored vehicle.
3. The vehicle information recording management method according to claim 2, wherein the obtaining the license plate profile of the monitoring vehicle by an edge detection algorithm according to the aspect ratio and the color distribution includes:
performing binarization processing on the acquired image to obtain a binarized image of the acquired image;
obtaining a plurality of geometric features of the monitoring vehicle through findcontours functions according to the binarized image;
obtaining an aspect ratio of each of the geometric features;
and screening the geometric features according to the aspect ratio, the aspect ratio of the geometric features and the color distribution to obtain the license plate outline.
4. The vehicle information recording management method according to claim 3, wherein the step of screening the geometric features to obtain the license plate outline according to the aspect ratio, the aspect ratio of the geometric features, and the color distribution includes:
Obtaining the corresponding color of the geometric feature in the acquired image according to the color distribution;
screening the geometric features according to the aspect ratio, the preset license plate aspect ratio and the preset color;
and judging the outline of the geometric feature as the license plate outline when the difference between the aspect ratio of the geometric feature and the aspect ratio of the preset license plate is smaller than a preset threshold value and the color difference between the color and the preset color is smaller than a preset color difference threshold value.
5. The vehicle information recording management method according to claim 3, wherein the processing the license plate contour through morphological operations according to the license plate contour to obtain the enhanced image of the license plate contour includes:
dividing the area where the license plate outline is located in the binarized image to obtain a license plate binarized image;
Performing expansion operation on the license plate binarization image to obtain an expanded binarization image;
and performing corrosion operation on the expanded binary image to obtain a corroded binary image, and taking the corroded binary image as the enhanced image of the license plate outline.
6. The vehicle information recording management method according to claim 1, wherein the dividing the enhanced image of the license plate contour into a non-blocking area and a blocking area by a deep learning model according to the enhanced image of the license plate contour comprises:
inputting the enhanced image into the deep learning model, and obtaining local characteristics of the enhanced image through a convolution layer of the deep learning model;
inputting the local features into a pooling layer of the deep learning model, reducing the space dimension of the local features, and obtaining the feature representation of the enhanced image through an activation function layer;
optimizing the characteristic representation through an cross-ratio loss function to obtain the optimized characteristic representation;
And carrying out pixel classification on the enhanced image according to the optimized characteristic representation to obtain the non-shielding region and the shielding region.
7. The method for managing the recording of vehicle information according to claim 1, wherein the repairing by the context information according to the non-blocking area to obtain the blocking image of the blocking area comprises:
Acquiring context information of the non-occlusion region, wherein the context information comprises image information of a surrounding preset region of the non-occlusion region;
Obtaining the overall structural characteristics of the license plate outline and the characteristics of the non-shielding region according to the image information of the surrounding preset region of the non-shielding region;
inputting the integral structure characteristics and the characteristics of the non-shielding area into a self-encoder to generate an estimated image of the shielding area;
splicing the image of the non-shielding area with the estimated image, and judging whether the spliced image is available or not;
if yes, the estimated image is used as the shielding image of the shielding area.
8. The vehicle information recording management method according to claim 2, wherein the obtaining the license plate number of the monitoring vehicle according to the image of the non-blocking area and the blocking image includes:
splicing the image of the non-shielding area and the shielding image to obtain a clear image corresponding to the license plate contour;
performing character segmentation according to the clear image to obtain the geometric characteristics of each character in the clear image;
And identifying each character according to the geometric features to obtain the license plate number of the monitoring vehicle.
9. The vehicle information recording management method according to claim 2, wherein the recording of the license plate number and the vehicle type of the monitoring vehicle includes:
correlating the license plate number with the vehicle type to obtain a comprehensive data record of the monitored vehicle;
automatically generating a text file corresponding to the comprehensive data record through an automatic script, and recording the text file into a database.
10. A vehicle information recording management system, comprising:
The acquisition unit is used for acquiring an acquisition image of the monitoring vehicle;
The feature extraction unit is used for obtaining the vehicle features of the monitoring vehicle according to the acquired image, wherein the vehicle features comprise the vehicle type, the aspect ratio and the color distribution of the monitoring vehicle;
The contour extraction unit is used for obtaining the license plate contour of the monitoring vehicle through an edge detection algorithm according to the aspect ratio and the color distribution;
The image processing unit is used for processing the license plate contour through morphological operation according to the license plate contour to obtain an enhanced image of the license plate contour;
Dividing the enhanced image of the license plate contour into a non-shielding region and a shielding region through a deep learning model according to the enhanced image of the license plate contour;
The image restoration unit is used for restoring according to the non-shielding area through the context information to obtain a shielding image of the shielding area;
the identification unit is used for obtaining the license plate number of the monitoring vehicle according to the image of the non-shielding area and the shielding image;
and the data importing unit is used for inputting the license plate number and the vehicle type of the monitored vehicle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510181727.2A CN119672654B (en) | 2025-02-19 | 2025-02-19 | Vehicle information collection and management method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510181727.2A CN119672654B (en) | 2025-02-19 | 2025-02-19 | Vehicle information collection and management method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN119672654A true CN119672654A (en) | 2025-03-21 |
| CN119672654B CN119672654B (en) | 2025-06-24 |
Family
ID=94990360
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202510181727.2A Active CN119672654B (en) | 2025-02-19 | 2025-02-19 | Vehicle information collection and management method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119672654B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6553131B1 (en) * | 1999-09-15 | 2003-04-22 | Siemens Corporate Research, Inc. | License plate recognition with an intelligent camera |
| KR20130114037A (en) * | 2013-09-04 | 2013-10-16 | 엠엠피씨 주식회사 | Masking and recovering method of privacy region |
| CN104408931A (en) * | 2014-10-29 | 2015-03-11 | 合肥指南针电子科技有限责任公司 | Incomplete sign license plate identification system and method |
| CN107545239A (en) * | 2017-07-06 | 2018-01-05 | 南京理工大学 | A kind of deck detection method matched based on Car license recognition with vehicle characteristics |
| CN115689946A (en) * | 2022-12-29 | 2023-02-03 | 北京集度科技有限公司 | Image restoration method, electronic device and computer program product |
| CN117115445A (en) * | 2023-09-14 | 2023-11-24 | 沈阳农业大学 | An image non-visible area completion method based on non-modal instance segmentation |
| CN117218599A (en) * | 2023-09-28 | 2023-12-12 | 深邦智能科技集团(青岛)有限公司 | Vehicle identification method and system for deliberately shielding license plate |
| CN118096593A (en) * | 2022-11-23 | 2024-05-28 | 海信集团控股股份有限公司 | Image processing method and electronic equipment |
| CN119314334A (en) * | 2024-12-13 | 2025-01-14 | 浪潮智慧科技有限公司 | A fee deduction method, system, device and medium for intelligent transportation system |
-
2025
- 2025-02-19 CN CN202510181727.2A patent/CN119672654B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6553131B1 (en) * | 1999-09-15 | 2003-04-22 | Siemens Corporate Research, Inc. | License plate recognition with an intelligent camera |
| KR20130114037A (en) * | 2013-09-04 | 2013-10-16 | 엠엠피씨 주식회사 | Masking and recovering method of privacy region |
| CN104408931A (en) * | 2014-10-29 | 2015-03-11 | 合肥指南针电子科技有限责任公司 | Incomplete sign license plate identification system and method |
| CN107545239A (en) * | 2017-07-06 | 2018-01-05 | 南京理工大学 | A kind of deck detection method matched based on Car license recognition with vehicle characteristics |
| CN118096593A (en) * | 2022-11-23 | 2024-05-28 | 海信集团控股股份有限公司 | Image processing method and electronic equipment |
| CN115689946A (en) * | 2022-12-29 | 2023-02-03 | 北京集度科技有限公司 | Image restoration method, electronic device and computer program product |
| CN117115445A (en) * | 2023-09-14 | 2023-11-24 | 沈阳农业大学 | An image non-visible area completion method based on non-modal instance segmentation |
| CN117218599A (en) * | 2023-09-28 | 2023-12-12 | 深邦智能科技集团(青岛)有限公司 | Vehicle identification method and system for deliberately shielding license plate |
| CN119314334A (en) * | 2024-12-13 | 2025-01-14 | 浪潮智慧科技有限公司 | A fee deduction method, system, device and medium for intelligent transportation system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN119672654B (en) | 2025-06-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107545239B (en) | Fake plate detection method based on license plate recognition and vehicle characteristic matching | |
| Ohya et al. | Recognizing characters in scene images | |
| Al-Ghaili et al. | Vertical-edge-based car-license-plate detection method | |
| CN103903005B (en) | License plate image identification system and method | |
| Antar et al. | Automatic number plate recognition of Saudi license car plates | |
| Khalifa et al. | Malaysian Vehicle License Plate Recognition. | |
| CN107895492A (en) | A kind of express highway intelligent analysis method based on conventional video | |
| Kaur | An automatic number plate recognition system under image processing | |
| CN111027535A (en) | License plate recognition method and related equipment | |
| CN112528917A (en) | Zebra crossing region identification method and device, electronic equipment and storage medium | |
| Shin et al. | Practical methods of image data preprocessing for enhancing the performance of deep learning based road crack detection | |
| Gilly et al. | A survey on license plate recognition systems | |
| CN112418100B (en) | Road vehicle abnormity detection method based on image classification and image comparison | |
| CN117576674A (en) | License plate recognition method, device, equipment and medium | |
| Paunwala et al. | A novel multiple license plate extraction technique for complex background in Indian traffic conditions | |
| Kročka et al. | Automatic license plate recognition using OpenCV | |
| CN110705553B (en) | A scratch detection method suitable for vehicle long-range images | |
| Tiwari et al. | Automatic vehicle number plate recognition system using matlab | |
| Sridevi et al. | Vehicle identification based on the model | |
| CN118366107B (en) | Irregular parking identification method on roads | |
| CN118196730B (en) | Method, device, equipment and storage medium for processing vehicle image data | |
| CN119672654B (en) | Vehicle information collection and management method and system | |
| CN117635581A (en) | Automobile VIN etching image positioning and defect detecting method | |
| CN116433927A (en) | Incomplete droplet identification method based on image feature detection algorithm | |
| Khan et al. | Advanced Techniques for Improved Bangladeshi Number Plate Detection and Character Recognition in Automated Parking Systems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |