CN114445992A - A real-time video surveillance method - Google Patents

Abstract

The invention discloses a real-time video monitoring method, which comprises collecting original video and filtering the sound in the original video to generate a processed video; decomposing the processed video frame by frame into a picture set, preprocessing the pictures, calculating Pixels at the edge of the image in the picture; detect the moving object in the video, and determine whether the frame of the picture is abnormal according to the position of the moving object; according to the number of frames in the abnormal picture, determine whether an alarm is required. In the process of real-time video recording, it can detect and judge the behavior of the target object in the video, make timely warning, and give timely warning to some violations, so as to prevent it from happening.

Description

A real-time video surveillance method

technical field

The invention relates to the technical field of real-time video monitoring, in particular to a real-time video monitoring method.

Background technique

In recent years, in today's society, with the rapid development of science and technology and the improvement of people's inner safety awareness, the use of monitoring systems in various fields of life is gradually popularizing, and it has been widely used in security, communication and transportation industries. Not only applied, but also gradually developed to other public industries. Video surveillance has been used in community supermarkets, railway stations, large shopping malls, banks and other related places where safety is critical. Routine maintenance can be made easier by consuming more human resources or money.

The video surveillance system is closely related to people's safety. If theft occurs, the police only need to recall the surveillance video of the day to find out what happened, and then they can reason and analyze based on the video surveillance, and finally solve the case successfully. The monitoring system can only record video, cannot detect and judge the behavior of the target in the video, and cannot make timely warnings, and it is difficult to give timely warnings to some violations. .

SUMMARY OF THE INVENTION

The purpose of this section is to outline some aspects of embodiments of the invention and to briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of the application to avoid obscuring the purpose of this section, abstract and title, and such simplifications or omissions may not be used to limit the scope of the invention.

The present invention has been proposed in view of the above-mentioned existing problems.

Therefore, the technical problem solved by the present invention is: the existing video monitoring system can only perform video recording, cannot detect and judge the behavior of the target object in the video, cannot make timely warning, and is difficult to give timely warning to some violations, It can only be traced back after the event, and it is difficult to prevent it before it happens.

In order to solve the above technical problems, the present invention provides the following technical solutions: collecting the original video and filtering the sound in the original video to generate a processed video; decomposing the processed video frame by frame into a picture set, preprocessing the pictures, calculating Pixels at the edge of the image in the picture; detect the moving object in the video, and determine whether the frame of the picture is abnormal according to the position of the moving object; according to the number of frames in the abnormal picture, determine whether an alarm is required.

As a preferred solution of the real-time video monitoring method of the present invention, the original video captured by the camera is collected, the original video is input into a computer, and the sound in the original video is filtered through a Gaussian filter.

As a preferred solution of the real-time video monitoring method of the present invention, wherein: the processed video is decomposed into picture sets frame by frame, the pictures in the picture set are smoothed by the Canny algorithm, and the The gradient strength and direction of the pixels at the edge of the image are expressed as:

Among them, σ represents the standard deviation, sigma=1.4, and e represents the pixel, then after Gaussian filtering, the value of the brightness of e is expressed as follows:

Among them, * represents the symbol of convolution, and sum represents the sum of all elements in the set matrix.

As a preferred solution of the real-time video monitoring method of the present invention, wherein: each direction of the pixel can be targeted by the edge in the image, and the four vertical, horizontal and diagonal edges of the image are detected by the Canny algorithm. direction. Through the operators representing the vertical and horizontal directions of edge detection, such as the first derivative values of G _x in the horizontal direction and G _y in the vertical direction, the direction theta of the pixel and the gradient G of the pixel can be inferred:

Among them, G represents the gradient strength, G _x represents the gradient amplitude in the x direction, G _y represents the gradient amplitude in the y direction, theta represents the direction of the pixel, arctan is the arc tangent function applied in the formula, and Sobel in the x and y directions The operators can be expressed as:

Among them, Sx represents the Sobel operator in the x direction, which is used to detect the direction where the edge direction is y, and Sy represents the Sobel operator in the y direction, which is used to detect the direction where the edge direction is x.

As a preferred solution of the real-time video monitoring method of the present invention, wherein: the non-maximum value suppression of edge pixels is compared to the gradient strength possessed by a single pixel, and the gradient strength of the current pixel at the same time and the other two Taking the pixel as an analogy, when the gradient strength of the current pixel is greater than the other two pixels, the current pixel is saved and the current pixel is used as the edge. The relational expression and the pseudocode related to non-maximum suppression are as follows:

Among them, tan is the tangent function, θ represents the angle, theta is the gradient direction of P, P represents the pixel, P1 represents the current pixel 1, and P2 represents the analog pixel.

As a preferred solution of the real-time video monitoring method of the present invention, wherein: edge pixels need to be filtered with weak gradient values, and some edge pixels are saved, and those with high gradient value properties, strong edge pixels have high gradient values. For the proposed high threshold, the gradient value of the weak edge pixel is smaller than the proposed high threshold and greater than the proposed low threshold. When the gradient value of the edge pixel is less than the proposed low threshold, the pixel will be suppressed, and the result is The pseudo code of double threshold detection, the specific expression is as follows:

As a preferred solution of the real-time video monitoring method according to the present invention, wherein: the position of the image of the moving object is different in the frames of the different images, and the position of the image of the moving object is different in the frames of the different images. Differences to calculate and analyze the changed pixels, based on the movement of the target changes with the change of time, reflecting strong continuity, the pixel changes of the background points in the picture tend to zero, and the processing is performed under the premise of threshold judgment. The frame numbers of the two subtracted frames are expressed as the kth frame and the (k+1)th frame, and the images of the frames can be expressed as I _k (x, y), I _k+1 (x, y), and the differential image is expressed as is D(x,y), and its binarization threshold is T, the corresponding formula of the inter-frame difference method is as follows:

Among them, I _x (x, y) represents the k-th frame image, I _k+1 (x, y) represents the (k+1)-th frame image, T is the binarization threshold, and D(x, y) represents a binary Value function, the value of 255 indicates that the pixels in the two frames of images have undergone significant changes. At this time, the pixel is located in the motion area. The value of 0 indicates that the pixel has almost no change, representing the background area. T stands for the established threshold.

As a preferred solution of the real-time video monitoring method of the present invention, the range of the threshold value T is used to adjust the sensitivity and accuracy of the detection target area. When the threshold value T is relatively small, some Noise or background is also mistakenly regarded as a moving target. When the threshold value T is large, a small part of the moving area will not be extracted.

As a preferred solution of the real-time video monitoring method of the present invention, wherein: when an abnormality occurs in the detection target area in the picture picture, the picture is defined as an abnormal picture, and the number of consecutive frames of the abnormal picture is preset. Occurrences greater than the number of frames trigger an alarm.

The beneficial effects of the invention are: in the process of real-time video recording, the behavior of the target object in the video can be detected and judged, and a timely warning can be given, and some illegal behaviors can be given a timely warning, which plays a preventive role.

Description of drawings

In order to illustrate the technical solutions of the embodiments of the present invention more clearly, the following briefly introduces the accompanying drawings used in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without any creative effort. in:

1 is a schematic flowchart of a basic flow of a real-time video monitoring method provided by an embodiment of the present invention;

2 is an image processing diagram of a real-time video monitoring method provided by an embodiment of the present invention;

FIG. 3 is an error comparison diagram of a real-time video monitoring method provided by an embodiment of the present invention.

Detailed ways

In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are a part of the embodiments of the present invention, not all of them. Example. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

Many specific details are set forth in the following description to facilitate a full understanding of the present invention, but the present invention can also be implemented in other ways different from those described herein, and those skilled in the art can do so without departing from the connotation of the present invention. Similar promotion, therefore, the present invention is not limited by the specific embodiments disclosed below.

Second, reference herein to "one embodiment" or "an embodiment" refers to a particular feature, structure, or characteristic that may be included in at least one implementation of the present invention. The appearances of "in one embodiment" in various places in this specification are not all referring to the same embodiment, nor are they separate or selectively mutually exclusive from other embodiments.

The present invention is described in detail with reference to the schematic diagrams. When describing the embodiments of the present invention in detail, for the convenience of explanation, the cross-sectional views showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the protection of the present invention. scope. In addition, the three-dimensional spatial dimensions of length, width and depth should be included in the actual production.

At the same time, in the description of the present invention, it should be noted that the orientation or positional relationship indicated in terms such as "upper, lower, inner and outer" is based on the orientation or positional relationship shown in the accompanying drawings, which is only for the convenience of describing the present invention. The invention and simplified description do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operate in a particular orientation, and therefore should not be construed as limiting the invention. Furthermore, the terms "first, second or third" are used for descriptive purposes only and should not be construed to indicate or imply relative importance.

Unless otherwise expressly specified and limited in the present invention, the term "installation, connection, connection" should be understood in a broad sense, for example: it may be a fixed connection, a detachable connection or an integral connection; it may also be a mechanical connection, an electrical connection or a direct connection. The connection can also be indirectly connected through an intermediate medium, or it can be the internal communication between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood in specific situations.

Example 1

Referring to FIG. 1, an embodiment of the present invention provides a real-time video monitoring method, including:

S1: Capture the original video and filter the sound in the original video to generate a processed video.

First of all, to realize the call of the camera, you must first install the hardware support package, which can be found in the additional functions under the start menu bar of matlab R2015a, otherwise you will not be able to use the computer's own camera or other cameras.

The "Additional Features Manager" page that opens has many devices, but we need to search for "camera" in the search box in the upper right corner, and then click Search to enter the search page.

After 15 results are found, click on the top title to enter the details page.

Click Install in the upper right corner, you will be prompted to select "Install" and "Download Only", select Install, and wait for a while.

After the installation is complete, find image acquisition under the app menu of matlab, click on it, and you can use the camera. After installation, you need to enter the image acquisition to use the camera.

Capture the original video shot by the camera, input the original video into the computer, and filter the sound in the original video through a Gaussian filter.

In the process of image analysis, it is the image quality that can affect the accuracy of the design and the effectiveness of the algorithm, so it needs to be preprocessed before the image is analyzed. Eliminating meaningless information, restoring meaningful information, improving the detection of effective information, and simplifying data to a great extent are the main purposes of image preprocessing, which can improve the reliability of subsequent image processing.

S2: Decompose the processed video frame by frame into a picture set, preprocess the picture, and calculate the pixels of the image edge in the picture.

It is the basic idea of this algorithm to find out the pixels with the individual maximum gradient magnitude in the image. Most tasks in detecting step edges converge on finding gradient numbers that approximate the image.

In the actual displayed image, the smoothing of the low-pass filter is inherent in the camera's optical system and circuit system, so its step edge is generally not very steep.

Using Gaussian filter to achieve the purpose of smoothing the image; calculating the magnitude and direction of the gradient, the method that can be used is the finite difference of the first-order partial derivative; using a non-maximum value, which can suppress the gradient magnitude; using double thresholds An algorithm that can perform detection while taking edge connections.

Smooth filtering through Canny algorithm, edge detection will bring different effects, in order to minimize the impact of noise on experimental data, it is necessary to carry out certain operations to filter out this noise, of course, it can also prevent the error detection. In order to smooth the processed image, convolution needs to be performed. The objects of convolution are the image and the Gaussian filter. After convolution, the image can be smoothed, and the influence of the noise of the edge detector can be greatly reduced. The processed video is decomposed into a picture set frame by frame, and the pictures in the picture set are smoothed by the Canny algorithm. The optimized part is to use the method of suppressing the isolated weak edge to realize edge detection in the last step, so that the obtained image edge contour is clearer , calculate the gradient intensity and direction of the pixels at the edge of the image in the image, and the expressions are as follows:

Among them, σ represents the standard deviation, sigma=1.4, and e represents the pixel, then after Gaussian filtering, the value of the brightness of e is expressed as follows:

Among them, * represents the symbol of convolution, and sum represents the sum of all elements in the set matrix.

Each direction of the pixel can be targeted by the edge in the image, and the four directions of vertical, horizontal and diagonal edges of the image are detected by the Canny algorithm. Through the operators representing the vertical and horizontal directions of edge detection, such as the first derivative values of G _x in the horizontal direction and G _y in the vertical direction, the direction theta of the pixel and the gradient G of the pixel can be inferred:

Among them, G represents the gradient strength, G _x represents the gradient amplitude in the x direction, G _y represents the gradient amplitude in the y direction, theta represents the direction of the pixel, arctan is the arc tangent function applied in the formula, and Sobel in the x and y directions The operators can be expressed as:

Among them, Sx represents the Sobel operator in the x direction, which is used to detect the direction where the edge direction is y, and Sy represents the Sobel operator in the y direction, which is used to detect the direction where the edge direction is x.

The non-maximum suppression of edge pixels is compared to the gradient strength of a single pixel, and the gradient strength of the current pixel is compared with the other two pixels at the same time. When the gradient strength of the current pixel is greater than the other two pixels, save the current pixel. pixel, using the current pixel as the edge,

Non-maximum suppression, its effect is reflected in the ability to "thin" sides. After performing related calculations on the image, such as gradient calculation, the edge extracted by the gradient value is still not very clear, and it is still somewhat blurred. In order to suppress all gradient values outside the local maximum to zero, the method of non-maximum suppression can be borrowed. The specific steps are as follows:

1) The one that needs to be compared is the gradient strength of the current pixel, and the other is the gradient strength of the two pixels along the positive and negative gradient directions.

2) Compare the gradient intensity of the current pixel with the other two pixels at the same time. When the current pixel is found to be the largest, the current pixel can be saved and used as an edge point. If the result is not the same, then the current pixel is suppressed. .

Generally, in order to obtain a more accurate calculation result, a linear interpolation method is used between two adjacent pixels beyond the gradient direction, so that the pixel gradients that need to be compared can be obtained.

After the above steps, the actual edge can be more accurately represented with the remaining pixels. However, there are still some edge pixels, which are caused by color and noise changes. Faced with these stray responses that need to be processed, edge pixels need to be filtered with weak gradient values, and some edge pixels with high gradient values are saved. Usually for this purpose, high and low thresholds are generally used. The definition of strong edge pixels is that the gradient value of edge pixels is higher than the proposed high threshold; weak edge pixels are defined as the gradient value of edge pixels is less than the proposed high threshold and greater than the proposed low threshold; when the gradient value of edge pixels Below the proposed low threshold, the pixel will be suppressed. The content of the input image can determine the choice of threshold, and the pseudocode related to the relational expression and non-maximum suppression is as follows:

Among them, tan is the tangent function, θ represents the angle, theta is the gradient direction of P, P represents the pixel, P1 represents the current pixel 1, and P2 represents the analog pixel.

S3: Detect the moving object in the video, and determine whether the picture of the frame is abnormal through the position of the moving object.

The background difference method can be applied to the background extraction algorithm of the still camera. It is necessary to select a still scene as the reference background. At this time, the pixels in the image all correspond to a value about the background, and the background value is relatively unchanged. The purpose of background selection is to select the background value of each point in the image sequence. Continuity is a feature of video sequences captured by cameras. When the moving target does not appear in the monitored scene, the continuous frames obtained at this moment will have an inconspicuous, almost faint change effect. But when moving objects appear, the constant frame-to-frame transitions become stark.

The position of the moving object is different in different frames of the image. The changed pixels are calculated and analyzed by the difference between the two adjacent frames around the picture set. The movement of the object changes with time. However, the change reflects a strong continuity. The pixel changes of the relevant background points in the picture tend to be zero, and the processing is carried out under the premise of threshold judgment. The frame numbers of the two subtracted frames are expressed as the kth frame and (k+th) 1) Frame, the image of the frame can be expressed as I _k (x, y), I _k+1 (x, y), and the differential image is expressed as D(x, y), and its binarization threshold is T, then The corresponding formula of the inter-frame difference method is as follows:

Among them, I _x (x, y) represents the k-th frame image, I _k+1 (x, y) represents the (k+1)-th frame image, T is the binarization threshold, and D(x, y) represents a binary Value function, the value of 255 indicates that the pixels in the two frames of images have undergone significant changes. At this time, the pixel is located in the motion area. The value of 0 indicates that the pixel has almost no change, representing the background area. T stands for The proposed threshold and the range of the threshold value T play a decisive role in the sensitivity and accuracy of detecting the target area. When the value of the threshold T is relatively small, some noises or backgrounds will be mistakenly regarded as moving targets. When the value of the threshold T is relatively large, a small part of the moving area will not be extracted. Choose an appropriate threshold T.

The range of the threshold value T is used to adjust the sensitivity and accuracy of the detection target area. When the threshold value T is relatively small, some noise or background will be mistakenly regarded as moving targets. When the threshold value T is large , a small part of the motion area is not extracted.

S4: According to the number of frames of abnormal pictures, determine whether an alarm is required. When there is an abnormality in the detection target area in the picture screen, define the picture as an abnormal picture, and preset the number of consecutive frames of abnormal pictures. When the number of abnormal pictures occurs is greater than The number of frames triggers the alarm. The inter-frame difference method usually selects two consecutive frames in time, or three consecutive frames as the images involved in the calculation. The overall advantage is that the principle is simple, the calculation method is simple, and the movement in the monitoring area can be detected in real time. of pixels.

Example 2

2 and 3 are another embodiment of the present invention, this embodiment is different from the first embodiment in that a verification test of a real-time video monitoring method is provided, in order to verify the technical effect adopted in the method and explain, In this embodiment, the traditional technical solution and the method of the present invention are used to carry out a comparative test, and the test results are compared by means of scientific demonstration, so as to verify the real effect of the method.

In terms of real-time video data collection, this experimental system uses the camera that comes with the notebook, and uses the videoinput function to open the system camera, uses the videoTimerFcn function to obtain the video stream, and uses the avp function to complete the real-time voice alarm function playback. In the detection of moving targets, the system adopts the inter-frame difference method, which is simple and easy to understand, suitable for moving targets, and can effectively improve the detection efficiency. The program is executed by a monolithic .m file. When an object passes through the camera, the surrounding of the figure interface turns red, and a voice alarm is issued in time, and the current picture is captured and sent to the mailbox of the emergency contact, indicating an abnormal state. If no target passes through the monitoring range of the camera, the screen around the figure is grayed out, indicating a normal state.

In order to ensure the sensitivity and accuracy of detecting the target area, we conducted multiple sets of experiments in order to obtain the optimal motion parameter threshold parameters, as shown in the following table:

Table 1: Table of mean error values for images on static backgrounds.

Table 2: Table of image mean error values under dynamic background.

Since the average error value of the image under the static background is less than 0.05, and the average error value of the image under the dynamic background is larger than 0.05, it is more appropriate to select 0.05 as the motion parameter threshold. At this time, the sensitivity and accuracy of detecting the target area are the best.

Experiments show that the motion parameter threshold is 0.05, the sound threshold parameter is 1.0, and the system works best when the frame number of each trigger image is 10. The system acquires the original image, adopts the Canny optimization algorithm to detect the edge of the original image, and then performs the inter-frame difference method. When the result is greater than a certain value, the surrounding of the real-time display screen turns red and an alarm is issued. On this basis, two groups of dynamic environment monitoring experiments are carried out. The results show that the real-time video monitoring system has a simple and clear operation interface, accurate It has the advantages of real-time, high security, and strong reliability. It can monitor the information of moving targets in the corresponding indoor area in real time, make timely warnings, and give timely warnings to some violations, which can prevent them from happening. effect.

It should be appreciated that embodiments of the present invention may be implemented or implemented by computer hardware, a combination of hardware and software, or by computer instructions stored in non-transitory computer readable memory. The methods can be implemented in a computer program using standard programming techniques - including a non-transitory computer-readable storage medium configured with a computer program, wherein the storage medium so configured causes the computer to operate in a specific and predefined manner - according to the specific Methods and figures described in the Examples. Each program may be implemented in a high-level procedural or object-oriented programming language to communicate with a computer system. However, if desired, the program can be implemented in assembly or machine language. In any case, the language can be a compiled or interpreted language. Furthermore, the program can be run on a programmed application specific integrated circuit for this purpose.

Furthermore, the operations of the processes described herein may be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The processes (or variations and/or combinations thereof) described herein can be performed under the control of one or more computer systems configured with executable instructions, and as code that executes collectively on one or more processors (eg, , executable instructions, one or more computer programs or one or more applications), implemented in hardware, or a combination thereof. The computer program includes a plurality of instructions executable by one or more processors.

Further, the methods may be implemented in any type of computing platform operably connected to a suitable, including but not limited to personal computer, minicomputer, mainframe, workstation, network or distributed computing environment, separate or integrated computers platform, or communicate with charged particle tools or other imaging devices, etc. Aspects of the invention may be implemented in machine-readable code stored on a non-transitory storage medium or device, whether removable or integrated into a computing platform, such as a hard disk, an optically read and/or written storage medium, RAM, ROM, etc., such that it can be read by a programmable computer, when a storage medium or device is read by a computer, it can be used to configure and operate the computer to perform the processes described herein. Furthermore, the machine-readable code, or portions thereof, may be transmitted over wired or wireless networks. The invention described herein includes these and other various types of non-transitory computer-readable storage media when such media includes instructions or programs that implement the steps described above in conjunction with a microprocessor or other data processor. The invention also includes the computer itself when programmed according to the methods and techniques described herein. A computer program can be applied to input data to perform the functions described herein, transforming the input data to generate output data for storage to non-volatile memory. The output information can also be applied to one or more output devices such as a display. In a preferred embodiment of the invention, the transformed data represents physical and tangible objects, including specific visual depictions of physical and tangible objects produced on the display.

As used in this application, the terms "component," "module," "system," etc. are intended to refer to a computer-related entity, which may be hardware, firmware, a combination of hardware and software, software, or running software. For example, a component can be, but is not limited to, a process running on a processor, a processor, an object, an executable, a thread in execution, a program, and/or a computer. As an example, both an application running on a computing device and the computing device may be components. One or more components can exist in a process and/or thread of execution, and a component can be localized on one computer and/or distributed between two or more computers. In addition, these components can execute from various computer readable media having various data structures thereon. These components can be implemented by, for example, having one or more data groupings (eg, data from one component interacting with another component in a local system, a distributed system, and/or in a signaling manner such as the Internet network to interact with other systems) to communicate locally and/or as remote processes.

It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be Modifications or equivalent substitutions without departing from the spirit and scope of the technical solutions of the present invention should be included in the scope of the claims of the present invention.

Claims (9)

1. A real-time video monitoring method, comprising:

collecting an original video and filtering sound in the original video to generate a processed video;

decomposing the processed video frame by frame into a picture set, preprocessing the picture, and calculating pixels at the edge of the picture;

detecting a moving object in the video, and judging whether the picture of the frame of picture is abnormal or not according to the position of the moving object;

and judging whether to alarm or not according to the number of the frames with the abnormal pictures.

2. The real-time video monitoring method of claim 1, wherein: the method comprises the steps of collecting an original video shot by a camera, inputting the original video into a computer, and filtering sound in the original video through a Gaussian filter.

3. The real-time video monitoring method of claim 2, wherein: decomposing the processed video frame by frame into a picture set, smoothing the pictures in the picture set by a Canny algorithm, and calculating the gradient strength and the direction of pixels at the edge of the image in the picture, wherein the expression is as follows:

where σ represents the standard deviation, sigma 1.4, and e represents the pixel, then the value of the luminance of e is expressed as follows after gaussian filtering:

where denotes the sign of the convolution and sum denotes the sum of all elements in the setting matrix.

4. The real-time video monitoring method of claim 3, wherein: each direction of the pixel may be represented by a graphThe edges in the image are aimed at, and the four directions of the vertical, horizontal and diagonal edges of the image are detected through the Canny algorithm. By operators representing vertical and horizontal directions of edge detection, e.g. G in horizontal direction_xAnd G in the vertical direction_yThe direction of the pixel theta and the gradient of the pixel G can be inferred by the resulting first derivative value:

wherein G represents gradient strength, G_xMagnitude of gradient, G, representing the x-direction_yRepresenting the magnitude of the gradient in the y-direction, theta representing the direction of the pixel, arctan being the arctan function applied in the formula, and Sobel operators in the x-and y-directions can be respectively expressed as:

wherein Sx represents a Sobel operator in the x direction for detecting the direction in which the edge direction is y, and Sy represents a Sobel operator in the y direction for detecting the direction in which the edge direction is x.

5. The real-time video monitoring method of claim 4, wherein: the non-maximum suppression of the edge pixel is compared with the gradient intensity of a single pixel, the gradient intensity of the current pixel is analogized with other two pixels at the same time, when the gradient intensity of the current pixel is greater than the gradient intensity of the other two pixels, the current pixel is stored, the current pixel is taken as the edge, and the relation expression and the pseudo code related to the non-maximum suppression are as follows:

G_P1＝(1-tan(θ))×E+tan(θ)×NE

G_p2＝(1-tan(θ))×W+tan(θ)×SW

if G_p≥G_p1 and G_p≥G_p2

G_p may be an edge

else

G_p should be sup pressed

where tan is a tangent function, θ represents an angle, theta is a gradient direction of P, P represents a pixel, P1 represents a current pixel 1, and P2 represents an analog pixel.

6. The real-time video monitoring method of claim 5, wherein: filtering the edge pixels by using weak gradient values, storing some edge pixels, wherein the edge pixels have the property of high gradient values, the gradient value of a strong edge pixel is higher than a proposed high threshold, the gradient value of a weak edge pixel is smaller than the proposed high threshold and larger than a proposed low threshold, and when the gradient value of the edge pixel is smaller than the proposed low threshold, the pixel is suppressed to obtain a pseudo code of double-threshold detection, and the specific expression is as follows:

if G_p≥HighThreshold

G_p is an strong edge

else if G_p≥LowThreshold

G_p is an weak edge

else

G_p should be sup pressed。

7. the real-time video monitoring method of claim 6, wherein: the images of moving objects are different in position in frames of different images, changed pixels are calculated and analyzed through the difference shown between two adjacent frames of images around a picture set, the motion of a target is changed along with the change of time, strong continuity is shown, the pixel change of a background point in the images tends to zero, processing is carried out on the premise of threshold judgment, the frame number of the two subtracted frames is shown as the kth frame and the (k +1) th frame, and the images of the frames can be shown as I_k(x,y)、I_k+1(x, y) Simultaneous Difference imageDenoted as D (x, y) and its binary threshold value is T, the corresponding formula of the inter-frame difference method is as follows:

wherein, I_x(x, y) denotes a k-th frame image, I_k+1The (x, y) represents the (k +1) th frame image, T is a binarization threshold, D (x, y) represents a binary function, the number of 255 indicates that pixel points in the two frame images are changed obviously, the pixel points are located in a motion area at the moment, the number of 0 indicates that the pixel points are almost not changed, the pixel points represent a background area, and T represents a well-formulated threshold.

8. The real-time video monitoring method of claim 7, wherein: the range size of the threshold value T is used for adjusting the sensitivity and accuracy of the detection target area, when the threshold value T is relatively small, certain noise or background can be mistakenly used as a moving target, and when the threshold value T is large, a small part in the range of the moving area can be not extracted.

9. The real-time video monitoring method of claim 8, wherein: when the abnormality occurs in the detection target area in the picture, defining the picture as an abnormal picture, presetting the number of frames in which the abnormal picture continuously occurs, and triggering an alarm when the occurrence frequency of the abnormal picture is more than the number of the frames.

Images