CN-121741134-B - Streaky pork fine division classification method, apparatus, electronic device and medium

CN121741134BCN 121741134 BCN121741134 BCN 121741134BCN-121741134-B

Abstract

The application relates to the technical field of meat segmentation, and provides a streaky pork fine segmentation classification method, a streaky pork fine segmentation classification device, electronic equipment and a medium, wherein the method comprises the following steps of shaping a sample to be detected; the method comprises the steps of shaping a sample to be detected, collecting an image frame of the sample to be detected, determining a sample interest area to be detected according to the sample image frame to be detected, establishing a sample characteristic binary template to be detected according to a muscle layer binary template and a fat layer binary template in the sample interest area to be detected, constructing and calculating a sample layer vector to be detected, and determining segmentation grading parameters according to element values in the sample layer vector to be detected. The method, the device and the electronic equipment provided by the application can accelerate the production efficiency, improve the automation degree and enlarge the production scale, so that the segmentation, classification and sorting links have self-adaptability, the accuracy of segmenting and classifying the high-quality parts of streaky pork is further improved, and the short plates of fresh meat in fine intelligent segmentation and classification at present are made up.

Inventors

WANG HUI
LI JIAPENG
WANG HAITANG
LI XIANG
ZOU HAO
WANG SHOUWEI
ZHAO YAN

Assignees

中国肉类食品综合研究中心

Dates

Publication Date: 20260505
Application Date: 20260227

Claims (9)

1. The streaky pork fine division and classification method is characterized by comprising the following steps of: Shaping a sample to be detected; Acquiring an image frame of the shaped sample to be detected, and determining an interest area of the sample to be detected according to the image frame of the sample to be detected; According to the muscle layer binary template and the fat layer binary template in the interest area of the sample to be detected, establishing a characteristic binary template of the sample to be detected, and constructing and calculating a layer vector of the sample to be detected; Determining segmentation classification parameters according to element values in the sample layer vector to be detected; Dividing, grading and sorting the sample to be detected according to the obtained dividing and grading parameters; The construction and calculation of the sample layer vector to be detected further comprises: Acquiring the Ferrett parameter of the binary template of the interest area of the sample to be detected; Rotating all pixel points in the sample interest area to be detected and the sample characteristic binary template to be detected so that the minimum Ferrett diameter is perpendicular to the X axis; constructing and calculating a layer vector of a sample characteristic binary template to be detected along the vertical direction; obtaining boundary pixels of the sample to be detected through the binary template of the sample interest area to be detected, and constructing an empty layer vector with the length being the number of the boundary pixels; And acquiring pixel values between each first and last pixel points on the boundary of the minimum Feret diameter direction of the sample characteristic binary template to be detected, sequentially carrying out derivation, accumulating the number of nonzero elements after derivation, adding 1, and sequentially filling the obtained values into the positions of the corresponding elements in the layer vectors.
2. The streaky pork fine division and classification method according to claim 1, wherein when the shaping operation step is performed, it is detected that the sample to be detected completely enters the shaping area, and the conveyor belt stops running; the shaping push rod pushes the shaping pressing plate to extrude the sample to be detected; Solving the average pressure value of a plurality of pressure sensors on the shaping pressing plate, and judging whether the average pressure value of the pressure sensors is larger than or equal to a preset pressure threshold value; when the pressure average value is greater than or equal to a preset pressure threshold value, the shaping push rod drives the shaping pressing plate to reset, the conveyor belt resumes operation, and the sample to be detected is conveyed to the detection area; When the pressure average value is smaller than a preset pressure threshold value, the shaping push rod pushes the shaping pressing plate to continuously squeeze the sample to be detected until the pressure average value of the response values of the pressure sensors is larger than or equal to the preset pressure threshold value.
3. The streaky pork fine division and classification method as set forth in claim 1, wherein the determination of the region of interest of the sample to be detected further includes the steps of: Acquiring an image frame of the sample to be detected at any angle; and obtaining a gray level image of the interest area of the sample to be detected and a binary template of the interest area of the sample to be detected, and judging whether the sample to be detected completely enters the field of view of the image analysis unit.
4. The streaky pork fine division and classification method as set forth in claim 3, characterized in that, Performing dynamic threshold segmentation on the gray level image of the interest region of the sample to be detected to respectively obtain a muscle layer binary template and a fat layer binary template; respectively constructing label matrixes of the muscle layer binary templates and the fat layer binary templates, sequentially filling holes of the two-value templates of each communicating object in the label matrixes of the muscle layer binary templates and the fat layer binary templates, and taking a difference set with the two-value templates of the original communicating object to obtain the two-value templates of intramuscular fat or connective tissue of the muscle layer and muscle bundles or connective tissue of the fat layer; and carrying out morphological opening operation on the binary templates of the intramuscular fat, the muscle bundles or the connective tissue by adopting a structure body with a preset size to obtain the characteristic binary templates of the sample to be detected.
5. The streaky pork fine division and classification method as claimed in claim 1, wherein, when determining the division and classification parameters based on the element values in the sample layer vector to be detected, Subtracting a preset layer number from the layer vector of the sample to be detected according to whether the sample to be detected is peeled or not; Sequentially deriving the difference vector values according to the difference between the element values in the sample layer vector to be detected and the preset layer number, obtaining a linear index of non-zero elements, adding 1, and marking as a segmentation position index to obtain an abscissa of a segmentation position of the sample to be detected; Judging the grade of each divided block according to the element value in the difference value vector between the divided part indexes; Dividing the binary templates of the interest areas of the samples to be detected according to the index of the dividing parts, establishing a region-separating label matrix of the interest areas of the samples to be detected, sequentially calculating the mass center coordinates of the binary templates of all the communicated objects in the matrix, and calculating the center coordinates of the central line of the mass center if the mass center is out of the binary templates of the communicated objects; And obtaining the time for the sample to be detected to reach the photoelectric sensor of the dividing and grading unit according to the value of the speed sensor of the detecting device, the distance from the first pixel of the boundary of the sample to be detected to the direction terminal of the detecting device and the distance from the sample transmitted by the dividing and grading unit to the photoelectric sensor.
6. The streaky pork fine division and classification method according to claim 1, wherein in the process of dividing, classifying and sorting the sample to be detected according to the obtained division and classification parameters, parameter matching is carried out according to the time when the sample to be detected arrives at the photoelectric sensor of the division and classification unit in the division and classification parameters of the sample to be detected, the time when the photoelectric sensor of the image analysis unit sends a transmission signal to the division and classification unit to detect that the sample to be detected arrives at a preset position, if the difference value of the two signals is larger than a preset threshold value, the next signal is waited, if the difference value of the two signals is smaller than the preset threshold value, the division and classification parameters are sent to a communication port, a conveyor belt of the division and classification unit is stopped, and a cutter sequentially moves to the transverse coordinate of a division part to cut the sample; obtaining the time of each divided block reaching the sorting wobble plate according to the numerical value provided by the speed sensor of the dividing and grading unit and the barycenter or central axis center abscissa of each divided block in the dividing and grading parameters; according to the grades of the dividing blocks in the parameters, after the dividing blocks reach the countdown zeroing of the sorting wobble plate, the sorting wobble plate turns to a sorting slideway with the corresponding grade.
7. A streaky pork fine division classifying device characterized in that a streaky pork fine division classifying method according to any one of claims 1 to 6 is applied, comprising: the sample shaping unit is used for straightening and flattening the shape of the manually-segmented sample to be detected and shaping the sample into a regular shape; the image analysis unit is used for acquiring image frames in the sample video stream to be detected in the transmission process under a preset angle; the parameter calculation unit is used for obtaining parameters required by segmentation, classification and sorting of the sample to be detected; the dividing and grading unit is used for conveying, dividing and sorting the samples to be detected.
8. An electronic device comprising a receptacle, a processor and a computer program stored on the receptacle and executable on the processor, characterized in that the processor implements the steps of the streaky pork fine segmentation classification method according to any one of claims 1-6 when the computer program is executed.
9. A non-transitory computer readable storage medium, having stored thereon a computer program which, when executed by a processor, implements the steps of the streaky pork fine division classification method as claimed in any one of claims 1 to 6.

Description

Streaky pork fine division classification method, apparatus, electronic device and medium Technical Field The application relates to the technical field of meat segmentation, in particular to a streaky pork fine segmentation classification method, a streaky pork fine segmentation classification device, electronic equipment and a medium. Background The streaky pork is mainly used as cold fresh raw material meat, a part of the streaky pork is used for being segmented and retail on the spot facing consumers at the terminal of a supermarket, a part of the streaky pork is used as modified atmosphere packaging segmented meat for being used as commercial finished products, and a part of the streaky pork is used for being processed into finished products or semi-finished products of various dishes by food enterprises. The method comprises the steps of determining fine segmentation sites among different layers of streaky pork by calculating and removing fat layers, muscle layers and interference factors in each layer of streaky pork, reserving complete and continuous fat texture, and mining high-quality parts in streaky pork products to the greatest extent. According to the labeling fine segmentation of the requirements of fat and lean cortex, the requirements of belt leather and the like, selling prices of different sales paths can be optimized. Currently, for the division, classification and sorting of streaky pork, each large-meat processing enterprise mainly adopts a manual mode, and in the division process, factors such as artificial subjective judgment, operation proficiency, processing intensity and the like are important influencing factors for influencing the stage sales of streaky pork. Therefore, a technology and a method for analyzing, dividing, classifying and sorting streaky pork products in real time in the production process are needed to replace or partially replace manual operation, enlarge the processing scale, improve the accuracy and the accuracy of dividing and classifying, improve the efficiency-to-cost ratio and further make up for the blank in the fine dividing and classifying and intelligent sorting of the streaky pork at present. Disclosure of Invention The application aims to provide a streaky pork fine division grading method, a streaky pork fine division grading device, electronic equipment and a medium, so as to solve or alleviate the problems in the prior art. In order to achieve the above object, the present application provides the following technical solutions: the application provides a streaky pork fine segmentation classification method, which comprises the following steps: the streaky pork fine division and classification method is characterized by comprising the following steps of: Shaping a sample to be detected; Acquiring an image frame of the shaped sample to be detected, and determining an interest area of the sample to be detected according to the image frame of the sample to be detected; According to the muscle layer binary template and the fat layer binary template in the interest area of the sample to be detected, establishing a characteristic binary template of the sample to be detected, and constructing and calculating a layer vector of the sample to be detected; Determining segmentation classification parameters according to element values in the sample layer vector to be detected; And dividing, grading and sorting the sample to be detected according to the obtained dividing and grading parameters. Optionally, when the shaping operation step is performed, detecting that the sample to be detected completely enters the shaping area, and stopping the operation of the conveyor belt; the shaping push rod pushes the shaping pressing plate to extrude the sample to be detected; Solving the average pressure value of a plurality of pressure sensors on the shaping pressing plate, and judging whether the average pressure value of the pressure sensors is larger than or equal to a preset pressure threshold value; when the pressure average value is greater than or equal to a preset pressure threshold value, the shaping push rod drives the shaping pressing plate to reset, the conveyor belt resumes operation, and the sample to be detected is conveyed to the detection area; When the pressure average value is smaller than a preset pressure threshold value, the shaping push rod pushes the shaping pressing plate to continuously squeeze the sample to be detected until the pressure average value of the response values of the pressure sensors is larger than or equal to the preset pressure threshold value. Optionally, determining the region of interest of the sample to be detected further comprises the steps of: Acquiring an image frame of the sample to be detected at any angle; and obtaining a gray level image of the interest area of the sample to be detected and a binary template of the interest area of the sample to be detected, and judging whether the sample to be detected completely en