CN117636334B - Pork quality classification method based on image processing - Google Patents
Pork quality classification method based on image processing Download PDFInfo
- Publication number
- CN117636334B CN117636334B CN202311750187.2A CN202311750187A CN117636334B CN 117636334 B CN117636334 B CN 117636334B CN 202311750187 A CN202311750187 A CN 202311750187A CN 117636334 B CN117636334 B CN 117636334B
- Authority
- CN
- China
- Prior art keywords
- pork
- image
- cover
- outer cover
- quality
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/68—Food, e.g. fruit or vegetables
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
- G06V10/12—Details of acquisition arrangements; Constructional details thereof
- G06V10/14—Optical characteristics of the device performing the acquisition or on the illumination arrangements
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
- G06V10/12—Details of acquisition arrangements; Constructional details thereof
- G06V10/14—Optical characteristics of the device performing the acquisition or on the illumination arrangements
- G06V10/141—Control of illumination
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Image Analysis (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
A pork quality classification method based on image processing relates to the field of pork quality classification, and comprises the following steps: s101, obtaining pork sample pictures, obtaining pictures with consistent standards through an image collector, detecting pork samples, and marking according to the quality of the pork samples according to detection results, wherein the pork samples are classified into good and poor; s102, preprocessing an image; s103, alexNet, training the model, and training the image by adopting a AlexNet model to obtain a trained classification model; s104, applying the trained classification model to pork quality classification. The pork quality classifying device has the beneficial effects that the model formed by training the pork sample image by adopting the AlexNet model can be used for rapidly classifying the pork quality.
Description
Technical Field
The invention relates to the field of pork quality classification, in particular to a pork quality classification method based on image processing.
Background
Pork, one of the most common meats worldwide, has received increasing attention from consumers and producers in its diversity and classification of quality. With the improvement of life quality, the demand of consumers for pork is gradually changed from pure quantity satisfaction to pursue quality and taste. This change has driven the pork industry to transform and upgrade, and has also made the pork quality classification a hotspot for industry research.
Currently, the traditional pork quality classification method mainly depends on the appearance, color, texture, fat content and the like of pork. For example, pork can be classified into streaky pork, lean meat, fat meat, etc. according to the lean degree of pork; according to the parts, it can be classified into pork neck, pork shoulder, pork tripe, etc. The sorting method is simple and visual, and is convenient for consumers to select, but has the defect of being too extensive and can not accurately reflect the inherent quality of pork. With the advancement of technology, more modern technology is being applied to pork quality classification. For example, the moisture, protein and fat contents of pork can be detected by utilizing near infrared spectrum technology, so that the pork can be more accurately classified; through the machine vision technology, the characteristics of pork such as color, texture, shape and the like can be automatically identified, and quick classification is realized; there are also DNA-based molecular marker technologies that can be used to classify pork quality from a genetic perspective.
With the development of technology, pork quality classification is becoming finer and more technological. On one hand, as consumers pay more attention to food safety and health, the requirements on pork quality become higher and higher; on the other hand, technological advances will provide more possibilities for pork quality classification. For example, the production process of pork is monitored and predicted in real time by using artificial intelligence and big data technology so as to realize more accurate quality control; or the pig breeds are improved by utilizing a gene editing technology so as to meet specific quality requirements.
In order to rapidly and objectively evaluate pork quality, various research institutions or companies have conducted intensive studies, such as: the publication number is CN101975844A, a pork quality comprehensive detection method based on a multi-sensor fusion technology is disclosed, and the method comprises the following steps: s1, detecting more than four indexes of an effective attenuation coefficient, an impedance spectrum, a sensory score, myooxygen saturation, a pH value, a chromaticity parameter and a bacterial colony total number of a pork sample to be classified, and marking the classification; s2, carrying out data fusion calculation and classification according to index parameters of the pork samples marked with the categories, and obtaining quality classification results of the pork samples. The method can accurately and rapidly carry out comprehensive evaluation on pork quality.
The method and the device for evaluating the quality of the live pork with high reliability are disclosed by the publication number CN116663771A, the situation that all parts of data cannot be tampered is realized based on a blockchain technology, the reliability is high, the quality of the live pork is simply evaluated according to the traceability information of the live pork, the quick identification of the live pork quality by purchasing personnel is facilitated, meanwhile, the simple classification of various pork types by a dealer is facilitated, the safety problem of the pork quality can be comprehensively controlled through the retrieval of the traceability information, the operation is simple, and the applicability is good.
The above methods all evaluate and classify pork quality indexes, but have certain subjective judgment, and cannot acquire pork sample images with consistent standards, and can not acquire more accurate pork quality classification only in terms of image processing.
Disclosure of Invention
The application aims to provide a pork quality classification method based on image processing, which is mainly used for classifying pork quality by processing pictures of pork quality samples and training AlexNet models.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
The pork quality classification method based on image processing is characterized by comprising the following steps:
S101, obtaining pork sample pictures, obtaining pictures with consistent standards through an image collector, detecting pork samples, and marking according to the quality of the pork samples according to detection results, wherein the pork samples are classified into good and poor;
s102, preprocessing an image;
s103, alexNet, training the model, and training the image by adopting a AlexNet model to obtain a trained classification model;
s104, applying the trained classification model to pork quality classification;
The image collector in the step S101 mainly includes: the novel multifunctional portable electric power tool comprises a base (1), an outer cover (2), an inner cover (3) and a cover plate (4), wherein the outer cover (2) is arranged on the base (1), the inner cover (3) is arranged in the outer cover (2), and the cover plate (4) is arranged at the top ends of the outer cover (2) and the inner cover (3);
the detection indexes of the pork sample are as follows: PH, color, fat content, and moisture content.
Further, the step S102 includes the following steps:
s201: processing the image into RGB format, the image size is 224×224;
S202: the image layer of the red channel is processed, and the R channel (red channel) in the image processed in the step S201 is processed by adopting the following formula:
Wherein the method comprises the steps of Is the red channel value of the image,Is the final reserved red channel value; through the formula, the pixel value of the fat part in the picture can be set to be 0, so that the influence on the myoglobin of pork due to the existence of intramuscular fat is reduced;
s203: and synthesizing a new picture, and forming a new pork image by the treated R channel, the untreated G channel and the untreated B channel.
Further, the AlexNet models in step S103 include 5 convolution layers, 3 pooling layers and 3 full-connection layers, each convolution layer includes a convolution kernel, an offset, a ReLU activation function and a local response normalization module, where the 1 st, 2 nd and 5 th convolution layers are all followed by one pooling layer, the last three layers are full-connection layers, and the final output layer is softmax, and the network output is converted into a probability value for predicting the class of the image.
Further, base (1) mainly includes pedestal (1-1), pedestal (1-1) is the bench shape, the top of pedestal (1-1) is equipped with dustcoat draw-in groove (1-2) for install dustcoat (2), the inboard of dustcoat draw-in groove (1-2) is equipped with seat bench ring (1-4), the top of seat bench ring (1-4) is equipped with LED lamp (1-5) for shoot pork sample image and provide the light source.
Further, an inner ring (1-3) is arranged on the inner side of the seat ring (1-4), an inner cover (3) is arranged on the inner side of the inner ring (1-3), a pork sample is placed at the central position of the inner ring (1-3), a cleaning groove (1-6) for removing dust or sundries in the outer cover clamping groove (1-2) is formed in the left side of the outer cover clamping groove (1-2), and a knob (1-7) for adjusting the brightness of the LED lamp (1-5) is arranged on the inclined surface of the seat body (1-1).
Further, apron (4) mainly includes lid (4-1), the central part of lid (4-1) is provided with through-hole (4-3), the outside of through-hole (4-3) is equipped with shallow slot (4-2) that suits with the taking lens for taking lens card is in shallow slot (4-2), the lower surface of lid (4-1) is equipped with cover ring (4-4) that suits with dustcoat (2) and inner cover (3) installation clearance, thereby can be with apron (4) fixed mounting on dustcoat (2) and inner cover (3).
Further, the pork sample is obtained by collecting pork with a thickness of 05cm-1.0cm in the middle section of the longest muscle of the lumbar vertebra back of the pig.
The beneficial effects are that:
1. the image angle and illumination intensity of the pork quality obtained by the image collector are consistent, so that the high-quality pork image can be obtained;
2. collecting pork from the eye muscle fixing part, and representing the quality of the pork through the eye muscle;
3. the pixels in the red channel domain value are averaged, so that interference pixels can be effectively eliminated, and effective utilization of pixel values is improved;
4. the model formed by training the pork sample image by adopting the AlexNet model can be used for quickly classifying the pork quality.
Drawings
FIG. 1 is a block flow diagram of a pork quality classification method based on image processing
FIG. 2 is a block diagram of an image preprocessing flow;
FIG. 3 is a diagram of the model structure of AlexNet models;
FIG. 4 is a schematic structural view of an image collector;
FIG. 5 is an exploded view of FIG. 4;
FIG. 6 is a schematic view of the structure of the base (1);
FIG. 7 is a schematic view of the structure of the cover plate (4);
fig. 8 is a top view of fig. 7.
1: Base, 2: outer cover, 3: inner cover, 4: cover plate, 1-1: base, 1-2: cover draw-in groove, 1-3: inner ring, 1-4: seat ring, 1-5: LED lamp, 1-6: cleaning groove, 1-7: knob, 4-1: cover body, 4-2: shallow groove, 4-3: through hole, 4-4: and (3) a cover ring.
Detailed Description
The invention provides a pork quality classification method based on image processing, which comprises the following steps as shown in fig. 1:
S101, obtaining pork sample pictures, obtaining pictures with consistent standards through an image collector, detecting pork samples, and marking according to the quality of the pork samples according to detection results, wherein the pork samples are classified into good and poor;
The indexes for detecting pork samples are respectively as follows: the pH value, the color and the fat content (the fat content is measured by adopting a extraction method, the extraction method is to dissolve and extract fat in a sample by utilizing an organic solvent (such as diethyl ether, petroleum ether and the like), the fat content is measured by weighing extract or residues), and the water content (the water content is measured by adopting a drying method, which is to measure the water content by utilizing the principle that water volatilizes at high temperature and by weighing the mass difference of the sample before and after drying).
Wherein the PH value of the pork with good quality is 5.8-6.2, and the pork with poor quality is out of the above range; moisture content: the moisture content of PSE meat is typically between 72% and 76%, the moisture content of DFD meat is typically between 58% and 62%, and the moisture content of normal meat is typically between 66% and 70%.
The color and luster are detected by a color difference meter, the L value of pork is generally between 40 and 60, the a value is between 10 and 20, the b value is between 5 and 15, and the meat quality in the range is good. When the L value is less than 40, the meat color is too dark; when the L value is higher than 60, the meat color is excessively pale. When the value a is lower than 10, the meat color is whitened; when the value a is higher than 20, the flesh color is red. When the value b is lower than 5, the meat color is blue; when the b value is higher than 15, the meat color is yellowish.
PSE meat typically has a fat content of between 2% and 4%, DFD meat typically has a fat content of between 4% and 6%, and normal meat typically has a fat content of between 3% and 5%.
It is good that the indexes are all in the normal range, and if one (including one) of the indexes is out of the normal range, the pork quality is poor.
S102, preprocessing an image;
As shown in fig. 2, the specific steps are as follows:
s201: processing the image into RGB format, the image size is 224×224;
S202: the image layer of the red channel is processed, and the R channel (red channel) in the image processed in the step S201 is processed by adopting the following formula:
Wherein the method comprises the steps of Is the red channel value of the image,Is the final reserved red channel value; through the formula, the pixel value of the fat part in the picture can be set to be 0, so that the influence on the pork myoglobin caused by the existence of intramuscular fat is reduced.
S203: and synthesizing a new picture, and forming a new pork image by the treated R channel, the untreated G channel and the untreated B channel.
S103, alexNet, training the model, and training the image by adopting a AlexNet model to obtain a trained classification model;
As shown in fig. 3, alexNet total convolutional layers (including 3 pooling) and 3 fully-concatenated layers. Wherein each convolution layer contains a convolution kernel, bias terms, a ReLU activation function, and a Local Response Normalization (LRN) module. The 1 st, 2 nd and 5 th convolution layers are all followed by a maximum pooling layer, and the last three layers are all connected layers. The final output layer is softmax, which converts the network output into a probability value for predicting the class of the image.
The AlexNet convolutional layers are specifically as follows:
Convolution layer C1: the input image of 224 x 3 is filtered using 96 kernels of 11 x 3 convolution kernel size with a step size of 4. A pair of 55×55×48 feature maps are placed in the ReLU activation function, respectively, to generate an activation map. The activated image is maximally pooled, the size is 3×3, the stride is 2, and the pooled feature map size is 27×27×48 (a pair). The LRN processing is performed after pooling.
Convolution layer C2: the output of convolution layer C1 (response normalization and pooling) is used as input and filtered using 256 convolution kernels, which are 5 x 48 in size.
Convolution layer C3: there are 384 kernels of size 3 x 256 connected to the output (normalized, pooled) of the convolutional layer C2.
Convolution layer C4: there are 384 cores, the core size being 3×3×192.
Convolution layer C5: there are 256 cores, with a core size of 3 x 192. Convolutional layer C5 is more pooled than C3, C4 layers, again with a pooling kernel size of 3 x 3 and stride of 2.
The upper convolutional layers C3, C4, C5 are interconnected without an access pooling layer or normalization layer in between.
Wherein, the case of the full connection layer is as follows:
Full tie layer F6: 4096 neurons were set, then stimulated by a ReLU function and then Dropout processed. The Dropout layer was used to reduce the occurrence of the overfitting phenomenon, and the Dropout layer had a parameter of 0.5.
Full tie layer F7: the same as the F6 layer.
Full tie layer F8: two neurons are arranged, and the activation function outputs two types of predicted values for softmax, and corresponds to the quality of pork.
S104, applying the trained classification model to pork quality classification.
The image collector in S101 mainly includes: base (1), dustcoat (2), inner cover (3) and apron (4), as shown in fig. 4 and 5, be equipped with dustcoat (2) on base (1), be equipped with inner cover (3) in the inside of dustcoat (2), be equipped with apron (4) on the top of dustcoat (2) and inner cover (3).
As shown in FIG. 6, the base (1) mainly comprises a base body (1-1), wherein the base body (1-1) is a stair platform, an outer cover clamping groove (1-2) is formed in the top end of the base body (1-1) and used for installing an outer cover (2), a base platform ring (1-4) is arranged on the inner side of the outer cover clamping groove (1-2), and an LED lamp (1-5) is arranged on the top end of the base platform ring (1-4) and used for providing a light source for shooting pork sample images. An inner ring (1-3) is arranged on the inner side of the seat ring (1-4), an inner cover (3) is arranged on the inner side of the inner ring (1-3), and a pork sample is placed in the central position of the inner ring (1-3). The left side of the outer cover clamping groove (1-2) is provided with a cleaning groove (1-6) for cleaning dust or sundries in the outer cover clamping groove (1-2). The inclined plane of the seat body (1-1) is provided with a knob (1-7) for adjusting the brightness of the LED lamp (1-5).
As shown in fig. 7 and 8, the cover plate (4) mainly comprises a cover body (4-1), a through hole (4-3) is formed in the center of the cover body (4-1), a shallow groove (4-2) which is suitable for a shooting lens is formed in the outer side of the through hole (4-3), the shooting lens is conveniently clamped in the shallow groove (4-2), and a cover ring (4-4) which is suitable for mounting gaps of the outer cover (2) and the inner cover (3) is arranged on the lower surface of the cover body (4-1), so that the cover plate (4) can be fixedly mounted on the outer cover (2) and the inner cover (3).
And acquiring a pork sample, namely acquiring pork with the thickness of 05cm-1.0cm of the middle section of the lumbar longus muscle (commonly called as eye muscle) of the pig, and after the acquisition is finished, putting the pork sample into an image acquisition device, and taking a pork picture through a camera.
Claims (4)
1. The pork quality classification method based on image processing is characterized by comprising the following steps:
S101, obtaining pork sample pictures, obtaining pictures with consistent standards through an image collector, detecting pork samples, and marking according to the quality of the pork samples according to detection results, wherein the pork samples are classified into good and poor;
s102, preprocessing an image;
s103, alexNet, training the model, and training the image by adopting a AlexNet model to obtain a trained classification model;
s104, applying the trained classification model to pork quality classification;
The image collector in the step S101 includes: the novel dust remover comprises a base (1), an outer cover (2), an inner cover (3) and a cover plate (4), wherein the outer cover (2) is arranged on the base (1), the inner cover (3) is arranged in the outer cover (2), the cover plate (4) is arranged at the top ends of the outer cover (2) and the inner cover (3), the base (1) comprises a base body (1-1), an outer cover clamping groove (1-2) is formed in the top end of the base body (1-1) and is used for installing the outer cover (2), and a cleaning groove (1-6) used for removing dust or sundries in the outer cover clamping groove (1-2) is formed in the left side of the outer cover clamping groove (1-2); the seat body (1-1) is in a trapezoid shape, a seat ring (1-4) is arranged on the inner side of the outer cover clamping groove (1-2), and an LED lamp (1-5) is arranged at the top end of the seat ring (1-4) and used for providing a light source for shooting pork sample images; an inner ring (1-3) is arranged on the inner side of the seat ring (1-4), the inner side of the inner ring (1-3) is used for installing an inner cover (3), a pork sample is placed at the central part of the inner ring (1-3), and a knob (1-7) for adjusting the brightness of the LED lamp (1-5) is arranged on the inclined surface of the seat body (1-1);
the detection indexes of the pork sample are as follows: PH, color, fat content, and moisture content;
the step S102 is as follows:
s201: processing the image into RGB format, the image size is 224×224;
s202: processing the image layer of the red channel, and processing the R channel in the image processed in the step S201 by adopting the following formula:
Wherein the method comprises the steps of Is the red channel value of the image,Is the final reserved red channel value; through the formula, the pixel value of the fat part in the picture is set to be 0, so that the influence on the myoglobin of the pork caused by the existence of the intramuscular fat is reduced;
s203: and synthesizing a new picture, and forming a new pork image by the treated R channel, the untreated G channel and the untreated B channel.
2. The method for classifying pork quality based on image processing according to claim 1, wherein: the AlexNet models in the step S103 include 5 convolution layers, 3 pooling layers and 3 full-connection layers, each convolution layer includes a convolution kernel, a bias term, a ReLU activation function and a local response normalization module, wherein the 1 st convolution layer, the 2 nd convolution layer and the 5 th convolution layer are all followed by one pooling layer, the last three layers are full-connection layers, and the final output layer is softmax, so that network output is converted into a probability value for predicting the type of an image.
3. The method for classifying pork quality based on image processing according to claim 2, wherein: the cover plate (4) comprises a cover body (4-1), a through hole (4-3) is formed in the center of the cover body (4-1), a shallow groove (4-2) which is adaptive to a shooting lens is formed in the outer side of the through hole (4-3) and used for clamping the shooting lens in the shallow groove (4-2), and a cover ring (4-4) which is adaptive to the mounting clearance of the outer cover (2) and the inner cover (3) is arranged on the lower surface of the cover body (4-1), so that the cover plate (4) is fixedly mounted on the outer cover (2) and the inner cover (3).
4. A method for classifying pork quality based on image processing according to claim 3, wherein: the pork sample is obtained by collecting pork with a thickness of 0.5cm-1.0cm in the middle section of the longest muscle of the lumbar vertebra back of the pig.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311750187.2A CN117636334B (en) | 2023-12-19 | 2023-12-19 | Pork quality classification method based on image processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311750187.2A CN117636334B (en) | 2023-12-19 | 2023-12-19 | Pork quality classification method based on image processing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117636334A CN117636334A (en) | 2024-03-01 |
| CN117636334B true CN117636334B (en) | 2024-07-26 |
Family
ID=90016352
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311750187.2A Active CN117636334B (en) | 2023-12-19 | 2023-12-19 | Pork quality classification method based on image processing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117636334B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118261492B (en) * | 2024-05-30 | 2024-08-20 | 绵阳梓兴食品科技有限公司 | Nondestructive testing method and system for comprehensive quality of fresh pork |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201780262U (en) * | 2010-09-09 | 2011-03-30 | 西南科技大学 | Pork external quality on-line grading device based on image processing |
| CN106485251A (en) * | 2016-10-08 | 2017-03-08 | 天津工业大学 | Egg embryo classification based on deep learning |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102323267A (en) * | 2011-08-10 | 2012-01-18 | 中国农业大学 | System and method used for rapidly evaluating freshness of raw meat products |
| CN117197559A (en) * | 2023-09-06 | 2023-12-08 | 郑州大学 | Pork classification model, construction method, electronic device and computer-readable medium based on deep learning |
-
2023
- 2023-12-19 CN CN202311750187.2A patent/CN117636334B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201780262U (en) * | 2010-09-09 | 2011-03-30 | 西南科技大学 | Pork external quality on-line grading device based on image processing |
| CN106485251A (en) * | 2016-10-08 | 2017-03-08 | 天津工业大学 | Egg embryo classification based on deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117636334A (en) | 2024-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Mendoza et al. | Colour and image texture analysis in classification of commercial potato chips | |
| CN103955938B (en) | A kind of wheat growth method for diagnosing status based on mobile Internet pattern and leaf color analysis | |
| Park et al. | Multispectral imaging system for fecal and ingesta detection on poultry carcasses | |
| CN101692052B (en) | Hyperspectrum image technique-based method and hyperspectrum image technique-based device for identifying true and false famous tea | |
| CN106525732B (en) | Rapid non-destructive testing method for internal and external quality of apples based on hyperspectral imaging technology | |
| CN102521564A (en) | Method for identifying tea leaves based on colors and shapes | |
| CN102854148A (en) | Detection and grading system for tenderness of fresh beef based on multispectral imagery | |
| CN105931223B (en) | The dividing method in maize portion in high spectrum reflection image based on wave band ratio method | |
| Pedreschi et al. | Computer vision classification of potato chips by color | |
| CN114419311B (en) | A method and device for nondestructive testing of passion fruit maturity based on multi-source information | |
| CN108613989A (en) | A kind of detection method of paddy unsound grain | |
| CN117636334B (en) | Pork quality classification method based on image processing | |
| Eissa et al. | Understanding color image processing by machine vision for biological materials | |
| US20080247603A1 (en) | Portable tool for determining meat quality | |
| CN106570855A (en) | Method and system for quickly judging pork freshness | |
| CN103076288A (en) | Automatic fish flesh grading device and method based on computer vision | |
| CN113340914B (en) | Method for measuring and calculating water eutrophication state based on hyperspectral remote sensing | |
| Tan et al. | Assessment of fresh pork color with color machine vision | |
| Ghasemi-Varnamkhasti et al. | Application of image analysis combined with computational expert approaches for shrimp freshness evaluation | |
| CN106383095A (en) | Device and method for detecting total number of bacteria on surface of cooled mutton | |
| CN102300100A (en) | White balance control device and method | |
| Asmara et al. | Chicken meat freshness identification using colors and textures feature | |
| CN117934962B (en) | Pork quality classification method based on reference color card image correction | |
| CN109089992A (en) | A kind of classification method and system of the Estimation of The Fish Freshness based on machine vision | |
| CN118097270A (en) | A soybean seed classification method based on RGB reconstruction of hyperspectral images |
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 |