Chicken carcass intelligent grading and segmenting system and method based on machine vision

Abstract

The invention relates to the technical field of chicken cutting, in particular to an intelligent chicken carcass grading and cutting system and method based on machine vision, and the system comprises a raw material conveying device, an industrial personal computer, a collection station, a removal station and a cutting station; the acquisition station comprises a camera, chicken carcasses are hung upside down on the raw material conveying device, chicken carcass images are acquired through the camera and uploaded to the industrial personal computer, and the industrial personal computer classifies the specifications of the chicken carcasses according to image information; and the industrial personal computer rejects the chicken carcasses which do not meet the requirements at the rejecting station according to the grading result of the chicken carcasses, and the chicken carcasses which meet the requirements continue to enter the cutting station along with the raw material conveying device to be cut. According to the invention, the chicken carcass length is graded to adapt to an automatic head cutting process, the body width is graded to adapt to an automatic wing cutting process, and the weight of a single chicken carcass is graded or the body width and the weight of the chicken carcass are graded at the same time to adapt to an automatic multi-part cutting process; therefore, the segmentation accuracy and the working efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of chicken cutting, and in particular to a chicken carcass intelligent grading and cutting system and method based on machine vision. Background Art

[0002] Chicken and its deep-processed products have the advantages of high nutrition, balanced protein and fat, and no religious dietary restrictions, and are deeply loved by consumers. In recent years, the consumption of chicken products has increased year by year. In order to meet the growing demand for chicken among consumers, the broiler processing process has gradually developed in the direction of automation, intelligence, efficiency, and cleanliness. my country's local chickens have strong adaptability and disease resistance, rich meat flavor, and are more in line with cooking needs and national eating habits. They occupy an important position in the chicken consumption market. However, compared with fast-growing large white-feathered broilers, local chicken carcasses have the characteristics of diversified varieties, large differentiation, and poor uniformity, making automatic segmentation technology a common problem in the industry, seriously hindering the development of efficient production in the local chicken industry. Summary of the invention

[0003] To solve the above problems, the present application provides a chicken carcass intelligent grading and segmentation system and method based on machine vision, which performs different types of grading on chicken carcasses and removes chicken carcasses that do not meet the requirements to facilitate segmentation of chicken carcasses in different parts, thereby improving segmentation efficiency.

[0004] The technical solution adopted by the present invention to solve the technical problem is: The present invention provides a chicken carcass intelligent grading and segmentation system based on machine vision, comprising a raw material conveying device, an industrial control computer, and a collection station, a rejection station and a segmentation station sequentially arranged along the raw material conveying direction; The collection station includes a camera and a background light board respectively located on both sides of the raw material conveying device. The chicken carcass is hung upside down on the raw material conveying device with the chest of the chicken carcass facing the camera. The camera collects the image information of the chicken carcass and uploads it to the industrial computer. The industrial computer grades the specifications of the chicken carcass according to the image information. The rejecting station includes a rejecting device and a material discharge chute respectively located on both sides of the raw material conveying device. A material discharge conveying device is arranged below the material discharge chute. The industrial computer controls the rejecting device to reject the chicken carcasses that do not meet the requirements according to the grading results of the chicken carcasses, and the chicken carcasses that meet the requirements continue to enter the cutting station with the raw material conveying device for cutting.

[0005] As an improvement of the above solution, the collecting station, the rejecting station and the dividing station are all provided with incoming material sensing devices.

[0006] As an improvement of the above solution, the industrial computer is connected to a display screen and a human-computer interaction touch panel.

[0007] As an improvement of the above solution, the number of the collection stations is 2, namely the first collection station and the second collection station in sequence.

[0008] As an improvement of the above scheme, the specifications of the chicken carcass include the length, body width and weight of the chicken carcass; the industrial computer obtains the length, body width and projected area of ​​the chicken carcass as characteristic parameters through the image information collected by the camera, inputs the characteristic parameters as a single or combined input, establishes a regression model for predicting the grade of the chicken carcass, and automatically determines the grade of the chicken carcass; wherein the weight of the chicken carcass is predicted by the three characteristic parameters of the length, body width and projected area of ​​the chicken carcass.

[0009] As an improvement of the above scheme, When the chicken carcass needs to be headed, the segmentation station includes an automated head-cutting device, and the industrial computer inputs the length of the chicken carcass as a single characteristic parameter, establishes a regression model for predicting the length grade of the chicken carcass, and automatically determines the length grade of the chicken carcass; When the chicken carcass needs to be wing-cut, the segmentation station includes an automated wing-cutting device, and the industrial computer inputs the body width of the chicken carcass as a single characteristic parameter, establishes a regression model for predicting the body width grade of the chicken carcass, and automatically determines the body width grade of the chicken carcass; When the chicken carcass is cut into multiple parts, the cutting station is an automated cutting system, and the industrial computer inputs the length, body width and projection area of ​​the chicken carcass as combined characteristic parameters, establishes a regression model for predicting the weight grade of the chicken carcass, and automatically determines the weight grade of the chicken carcass.

[0010] As an improvement of the above scheme, when the chicken carcass is divided into multiple parts, the industrial computer inputs the body width of the chicken carcass as a single characteristic parameter, establishes a regression model for predicting the body width grade of the chicken carcass, and automatically determines the body width grade of the chicken carcass; The industrial computer uses the length, body width and projection area of ​​the chicken carcass as combined characteristic parameters, establishes a regression model for predicting the weight grade of the chicken carcass, and automatically determines the weight grade of the chicken carcass; The chicken carcasses that meet the requirements are selected based on the weight and body width of the chicken carcasses and enter the cutting station for cutting into multiple parts.

[0011] The present invention also provides a method for intelligent grading and segmenting of chicken carcasses based on machine vision, which is applied to any of the intelligent grading and segmenting systems for chicken carcasses based on machine vision described in Shanshu, and the method comprises: Step 1: Hang the chicken carcass naturally with the breast side facing the camera; Step 2: The chicken carcass is moved to the collection station to collect images of the chicken carcass, and the image information is uploaded to the industrial computer; Step 3: the industrial computer analyzes and processes the image information, selects required feature parameters, establishes a regression model for predicting the chicken carcass grade, and automatically determines the chicken carcass grade; Step 4: According to the grade results, the chicken carcasses that do not meet the requirements are removed, and the remaining chicken carcasses continue to move to the cutting station; Step 5: Cut the chicken carcass into corresponding parts at the cutting station.

[0012] As an improvement of the above scheme, When the chicken carcass needs to be decapitated, In step 3, the length of the chicken carcass is selected as a single characteristic parameter input, a regression model for predicting the length grade of the chicken carcass is established, and the length grade of the chicken carcass is automatically determined; In step 5, the segmentation station is an automated head cutting device, and the head of the chicken carcass is segmented by the automated head cutting device; When the wing cutting is required for chicken carcasses, In step 3, the body width of the chicken carcass is selected as a single characteristic parameter input, a regression model for predicting the body width grade of the chicken carcass is established, and the body width grade is automatically determined; In step 5, the segmentation station is an automated wing cutting device, and the wings of the chicken carcass are segmented by the automated wing cutting device; When cutting a chicken carcass into multiple parts, In step 3, the length, body width and projected area of ​​the chicken carcass are selected as combined characteristic parameter inputs, a regression model for predicting the weight grade of the chicken carcass is established, and the weight grade is automatically determined; In step 5, the cutting station is an automated cutting system, through which the wings, legs and breasts of the chicken carcass are cut.

[0013] As an improvement of the above scheme, the regression model for length grade prediction is L1=0.211L-4.521, wherein L1 is the predicted value of the actual length of the chicken carcass, and L is the characteristic value of the chicken carcass length; The regression model for predicting the body width grade is W1=0.132W+2.131, wherein W1 is the predicted value of the actual body width of the chicken carcass, and W is the characteristic value of the body width of the chicken carcass; The regression model for weight grade prediction is M=0.423W+0.001S+0.017L-901.644, wherein M is the predicted value of the actual weight of the chicken carcass, W is the characteristic value of the width of the chicken carcass, S is the characteristic value of the projected area of ​​the chicken carcass, and L is the characteristic value of the length of the chicken carcass.

[0014] The beneficial effects brought by the present invention are: The present invention provides a machine vision-based intelligent grading and segmentation system and method for chicken carcasses. When a chicken carcass is hung on a raw material conveying device and passes through a collection station, a camera is used to collect image information of the chicken carcass, and the length, body width and projection area of ​​the chicken carcass are extracted. After processing and analyzing the obtained data, these characteristic parameters are used as single or combined inputs to establish a regression model for predicting the grade of the chicken carcass, and the grade of the chicken carcass is automatically determined. Chicken carcasses that do not meet the requirements after grading are removed at a removal station to meet the production needs of subsequent segmentation stations. The length of the chicken carcass is graded to adapt to the automated head cutting process, the body width is graded to adapt to the automated wing cutting process, and the weight of a single chicken carcass is graded or the body width and the weight of the chicken carcass are graded simultaneously to adapt to the automated multi-part segmentation process, thereby improving the segmentation accuracy and work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments. Attached Figure 1 This is a schematic diagram of the structure of the machine vision-based chicken carcass intelligent grading and segmentation system of the present application. Figure 1 ; Attached Figure 2 This is a schematic diagram of the structure of the machine vision-based chicken carcass intelligent grading and segmentation system of the present application. Figure 2 ; Attached Figure 3 This is a schematic diagram of the structure of the machine vision-based chicken carcass intelligent grading and segmentation system of the present application. Figure 3 .

[0016] In the figure: A. chicken carcass; 1. raw material conveying device; 11. slide rail; 12. hook; 2. first collection station; 21. camera; 22. background light board; 3. second collection station; 4. rejection station; 41. rejection cylinder; 42. unloading slide; 43. unloading conveying device; 44. rack guide rod; 5. control cabinet; 6. display screen; 7. human-computer interaction touch panel. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] In the description of the present invention, it should be understood that the terms "upper side", "lower side", "upper end", "two ends", "width", "height", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0019] In the present invention, unless otherwise clearly specified and limited, the terms "set", "install", "connect", "connect", "connection", "communication" and the like should be understood in a broad sense, for example, it can be a fixed connection or a detachable connection; it can be a mechanical connection; it can be directly connected or indirectly connected through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0020] Example 1 Reference Figure 1-3 , this embodiment provides a chicken carcass intelligent grading and segmentation system based on machine vision, comprising a raw material conveying device 1, an industrial computer, and a collection station, a rejection station 4 and a segmentation station sequentially arranged along the raw material conveying direction; The raw material conveying device 1 includes a slide rail 11 and a hook 12. The feet or legs of the chicken carcass A are hung on the hook 12. The chicken carcass A hangs naturally, with one side of the chicken breast facing the camera 21. The chicken carcass A moves along the slide rail 11 with the hook 12.

[0021] There are two collection stations, namely the first collection station 2 and the second collection station 3, which are arranged in sequence along the raw material conveying direction. The image information taken twice is uploaded to the industrial computer, which doubly guarantees the authenticity of the image information and improves the accuracy of grading.

[0022] The collection station includes a camera 21 and a background light panel 22 respectively located on both sides of the material conveying device 1. The background light panel 22 is two LED light source panels. The camera 21 is placed in a stainless steel protective box, and an opening adapted to the lens of the camera 21 is opened on the stainless steel protective box.

[0023] The input end of each collection station is provided with an incoming material sensing device, which is used to detect the chicken carcass A entering the collection station and send a detection signal to the camera 21, and the camera 21 takes a picture of the chicken carcass A according to the received detection signal.

[0024] The above-mentioned rejection station 4 includes a rejection device and a material discharge chute 42 respectively located on both sides of the raw material conveying device 1. The rejection device can be a rejection cylinder 41. A material discharge conveying device 43 is arranged below the material discharge chute 42. A material incoming sensing device is also arranged on one side of the rejection cylinder 41. When the current chicken carcass A does not meet the requirements, the industrial computer controls the rejection cylinder 41 to reject the chicken carcass A hung on the hook 12, and the rejected chicken carcass A slides along the material discharge chute 42 to the material discharge conveying device 43 for output. Specifically, the number of the rejection cylinders 41 is 3, and a material incoming sensing device is installed on one side of each rejection cylinder 41. The number of rejection cylinders 41 is controlled to start according to the number of chicken carcasses A that do not meet the requirements.

[0025] The above-mentioned rejection station 4 includes a hanger guide rod 44 arranged below the slide rail 11. The hanger guide rod 44 is lifted upward, so that the hook 12 is lifted upward along the hanger guide rod 44 at a certain angle, so that the rejection cylinder 41 can reject the chicken carcass A on the hook 12.

[0026] in, The incoming material sensing device is a fiber optic sensor.

[0027] The machine vision-based intelligent grading and segmentation system for chicken carcasses of this embodiment also includes a stainless steel frame, on which a raw material conveying device 1, a collection station, a rejection station 4, and a segmentation station are all installed. An industrial computer is arranged outside the frame, and the industrial computer is placed in a control cabinet 5, in which a cabinet air conditioner is also arranged. The industrial computer is connected to a display screen 6 and a human-computer interaction touch panel 7 for communication, and is equipped with a wireless keyboard and a mouse. The grading of the chicken carcass A can be viewed in real time through the industrial computer, and the images of the chicken carcass A taken in real time and the grading judgment result data can be stored and scheduled.

[0028] During operation, the raw material conveying device 1 transports the chicken carcass A to the collection station, the incoming material sensing device senses the chicken carcass A, the camera 21 takes a picture of the chicken carcass A, and transmits the image information to the industrial computer. The industrial computer performs real-time grading processing on the chicken carcass A through the image information, and outputs the grading result in a data queue. The industrial computer controls the rejecting cylinder 41 according to the grading result to remove the chicken carcasses A that do not meet the requirements, and the chicken carcasses A that meet the requirements continue to enter the cutting station with the raw material conveying device 1 for cutting at the corresponding position.

[0029] When cutting the heads of chicken carcass A, the segmentation station includes an automated cutting device. Since the position of the cutter in the automated cutting device for chicken carcass A is fixed, the entire chicken head must be below the cutter when the raw material enters the automated cutting device, otherwise the raw material will be damaged. Therefore, it is necessary to perform fast, accurate and automatic grading according to the length of chicken carcass A, which is a prerequisite for meeting the requirements of the automated cutting process.

[0030] The industrial computer uses the length of chicken carcass A as a parameter for grade judgment based on the image information of chicken carcass A, selects the extracted length of chicken carcass A as a single feature parameter input, and takes the predicted value of the estimated actual length of chicken carcass A as output. By comparing the predicted value of the length of chicken carcass A with the preset value of the grading length of chicken carcass A, the length grade of chicken carcass A is automatically determined, and the rejection station 4 sequentially rejects the chicken carcasses A that are not suitable for the size of the automated head cutting equipment, and the remaining chicken carcasses A enter the automated head cutting equipment to complete the head cutting process.

[0031] When the chicken carcass A is wing-cut, the cutting station includes an automatic wing-cutting device. When the chicken carcass A enters the automatic wing-cutting device in the cutting line, the slot width of the automatic wing-cutting device is fixed. When the raw material enters the automatic wing-cutting device, the body must be able to enter the slot. If the body is too large, the chicken carcass A cannot enter the wing-cutting machine. Therefore, it is necessary to perform fast and accurate automatic grading according to the body width, which is a prerequisite for meeting the automatic wing-cutting process.

[0032] The industrial computer uses the body width as a parameter for grade judgment based on the image information of the chicken carcass A, selects the extracted body width as a single feature parameter input, and uses the estimated predicted value of the body width as output. By comparing the predicted value of the body width with the preset value of the body width, the body width grade is automatically determined, and the rejection station 4 sequentially rejects the chicken carcasses A that are not suitable for the size of the automated wing cutting equipment, and the remaining chicken carcasses A enter the automated wing cutting equipment to complete the wing cutting process.

[0033] When the chicken carcass A is divided into multiple parts, the dividing station is an automated dividing system, which can be an automated dividing system produced by MEYN Company of the Netherlands, Marel Stork Poultry Processing Company and LINCO FOOD SYSTEM of Denmark. This online automated dividing system can automatically divide the chicken carcass A into multiple parts such as whole legs, three-section wings, and chicken breasts. However, due to the needs of the broiler processing production line and the inability of the equipment design to change its structure autonomously, the weight of the raw chicken carcass A needs to be controlled within a reasonable range of ±20% to achieve accurate and efficient automated dividing. Because chicken carcasses A, especially local chicken carcasses A, have large individual differences. Therefore, before entering the dividing station, the chicken carcass A needs to be quickly and accurately graded automatically according to its weight, which is the prerequisite for applying the automated dividing system.

[0034] According to the image information of the chicken carcass A, the industrial computer selects and extracts three characteristic parameters of the chicken carcass A, namely the length, body width and projection area, and takes the combination of these three parameters as input to construct a regression model of the predicted value of the chicken carcass A and the combination of the characteristic parameters, and takes the predicted value of the estimated weight of the chicken carcass A as output. By comparing the predicted value of the weight of the chicken carcass A with the preset value of the graded weight of the chicken carcass A, the weight grade of the chicken carcass A is automatically determined, and the chicken carcasses A that are not suitable for the automated cutting system are sequentially removed by the removal station 4, and the remaining chicken carcasses A can smoothly enter the automated cutting system.

[0035] in, The above-mentioned grading algorithms for chicken carcass A length, body width and chicken carcass A weight can all be completed using the VisionMaster 7100 Pro mature algorithm platform.

[0036] Example 2 This embodiment provides a method for intelligent grading and segmenting of chicken carcasses based on machine vision. The method is applied to the intelligent grading and segmenting system for chicken carcasses based on machine vision in the above-mentioned embodiment 1. The method is applied when the chicken carcass A needs to be headed, and the segmentation station includes an automated head-cutting device. The method includes: Step 1: Hang the feet of the chicken carcass A on the hook 12, let the chicken carcass A hang naturally, and face the camera 21 with the chicken breast side, and the chicken carcass A moves along the slide rail 11 with the hook 12; Step 2: After the fiber optic sensor of the first collection station 2 detects the chicken carcass A, it triggers the background light board 22 to flash and the camera 21 to take pictures, and uploads the image information to the industrial computer; The chicken carcass A enters the second collection station 3. After the optical fiber sensor of the second collection station 3 detects the chicken carcass A, it triggers the background light board 22 to flash and the camera 21 to take pictures, and uploads the image information to the industrial computer; Step 3: The industrial computer analyzes and processes the image information, selects the length of the chicken carcass A as a single feature parameter input, and calculates the predicted value of the actual length of the chicken carcass A and outputs the predicted value of the actual length of the chicken carcass A. By comparing the predicted value of the actual length of the chicken carcass A with the preset value of the grading length of the chicken carcass A, the length grade of the chicken carcass A is automatically determined, and the obtained grading result is displayed on the display screen 6; Step 4: Based on the classification results obtained in step 3, the chicken carcasses A that are not suitable for the size of the automatic head cutting equipment are sequentially removed by the removal cylinder 41; Step 5: The remaining chicken carcass A enters the automated head cutting equipment to complete the head cutting process.

[0037] in, It is known that in a certain 817 broiler slaughterhouse, the length of the chicken carcass A with the head and neck is in the range of 45.1-59.5cm, and for the automated head cutting equipment, the length of the chicken carcass A must be greater than or equal to 50cm to complete the automated head cutting process. Therefore, the preset value of the grading length L1 of the chicken carcass A ≥50cm is defined as "yes", and <50cm is defined as "no"; after comparing the length of the chicken carcass A, the grading result is output and displayed. The picking cylinder 41 picks out the chicken carcasses A with the output of "no" in turn, and the remaining chicken carcasses A can be automatically head cut smoothly. According to statistics from multiple independent tests, the accuracy of this embodiment using the characteristic parameters of the length of the chicken carcass A to predict the actual length of the chicken carcass A and output the grading result is over 98%.

[0038] Example 3 This embodiment provides a method for intelligent grading and segmenting of chicken carcasses based on machine vision. The method is applied to the intelligent grading and segmenting system for chicken carcasses based on machine vision in the above-mentioned embodiment 1. The method is applied when chicken carcass A needs to be wing cut. At this time, the segmentation station includes an automated wing cutting device. The method includes: Step 1: Hang the feet of the chicken carcass A on the hook 12, let the chicken carcass A hang naturally, and face the camera 21 with the chicken breast side, and the chicken carcass A moves along the slide rail 11 with the hook 12; Step 2: After the fiber optic sensor of the first collection station 2 detects the chicken carcass A, it triggers the background light board 22 to flash and the camera 21 to take pictures, and uploads the image information to the industrial computer; The chicken carcass A enters the second collection station 3. After the optical fiber sensor of the second collection station 3 detects the chicken carcass A, it triggers the background light board 22 to flash and the camera 21 to take pictures, and uploads the image information to the industrial computer; Step 3: The industrial computer analyzes and processes the image information, selects the body width as a single characteristic parameter input, and uses the pre-built regression model for predicting the body width grade: W1=0.132W+2.131 (R²=0.999, P<0.01), (W1 is the predicted value of the actual body width of the chicken carcass A, W is the characteristic value of the body width of the chicken carcass A, and the coefficient of the model is as high as 0.999, which can well predict the actual width of the chicken carcass), calculates the predicted value of the actual body width of the chicken carcass, and uses the predicted value of the actual body width of the chicken carcass as output, and automatically determines the body width grade by comparing the predicted value of the body width with the preset value of the chicken carcass grading width, and displays the obtained grading result on the display screen 6; Step 4: According to the classification result obtained in step 3, the chicken carcasses A that are not suitable for the size of the automatic wing cutting equipment are kicked out in turn by the ejection cylinder 41; Step 5: The remaining chicken carcass A enters the automated wing cutting equipment to complete the wing cutting process.

[0039] in, It is known that the body width range of a yellow-feathered broiler slaughterhouse is 10.3-17.4cm, and for the automated wing cutting equipment, the body width needs to be less than or equal to 16cm to smoothly enter the slot in the automated wing cutting equipment and complete the wing cutting process. Therefore, the preset value of body width W1 ≤16cm is defined as "yes", and >15cm is defined as "no"; after width comparison, the grading result is output and displayed. The ejection cylinder 41 kicks the chicken carcasses A with an output of "no" in turn, and the remaining chicken carcasses A can smoothly enter the automated wing cutting machine. According to statistics from multiple independent tests, the accuracy of this embodiment in predicting the actual width of the chicken carcass using the body width characteristic parameters and outputting the grading results is over 95%.

[0040] Example 4 This embodiment provides a method for intelligent grading and segmenting of chicken carcasses based on machine vision. The method is applied to the intelligent grading and segmenting system for chicken carcasses based on machine vision in the above-mentioned embodiment 1. When the method is applied to segmenting a chicken carcass A into multiple parts, the segmentation station includes an automated segmentation system. The method includes: Step 1: Hang the feet of the chicken carcass A on the hook 12, let the chicken carcass A hang naturally, and face the camera 21 with the chicken breast side, and the chicken carcass A moves along the slide rail 11 with the hook 12; Step 2: After the fiber optic sensor of the first collection station 2 detects the chicken carcass A, it triggers the background light board 22 to flash and the camera 21 to take pictures, and uploads the image information to the industrial computer; The chicken carcass A enters the second collection station 3. After the optical fiber sensor of the second collection station 3 detects the chicken carcass A, it triggers the background light board 22 to flash and the camera 21 to take pictures, and uploads the image information to the industrial computer; Step 3: According to the grading results obtained in step 3, the industrial computer analyzes and processes the image information, and automatically extracts three effective characteristic parameters of the chicken carcass A length, body width and projection area, and predicts the weight grade according to the pre-built regression model: M=0.423W+0.001S+0.017L-901.644 (R²=0.968, P<0.01), (M is the predicted value of the actual weight of the chicken carcass A, W is the body width of the chicken carcass A characteristic value, S is the characteristic value of the projection area of ​​the chicken carcass A, L is the characteristic value of the length of the chicken carcass A, the coefficient of the model is as high as 0.968, and the actual weight of the chicken carcass A can be well predicted), the predicted value of the actual weight of the chicken carcass A is calculated, and the predicted value of the weight of the chicken carcass A is used as output, and the weight grade of the chicken carcass A is automatically determined by comparing the predicted value of the weight of the chicken carcass A with the preset value of the graded weight of the chicken carcass A, and the obtained graded result is displayed on the display screen 6; Step 4: The ejection cylinder 41 sequentially ejects the chicken carcasses A that are not suitable for the weight of the automated cutting system; Step 5: The remaining chicken carcass A can smoothly enter the process-based automated cutting system to automatically cut the chicken carcass A into multiple parts such as the whole leg, three-section wing, and chicken breast.

[0041] in, It is known that the weight range of chicken carcass A in a local chicken slaughterhouse is 900-2200g / piece, and the standard value of the carcass weight that can be divided by the automatic cutting system purchased is set to 1500g. According to the weight range of chicken carcass A applicable to the system, the weight range fluctuates by 20%, and the preset value of the grading weight of chicken carcass A is set to 1200-1800g. About 75% of the raw chickens are within this range. Therefore, the preset value of the weight M of chicken carcass A <1200g is defined as "too small", >1800g is defined as "too large", and 1200g≤M≤1800g is defined as "appropriate"; after weight comparison, the grading result is output and displayed. The ejection cylinder 41 kicks the chicken carcasses A output as "too small" and "too large" in turn, and the remaining chicken carcasses A can smoothly enter the automated cutting system. After statistics of multiple independent tests, this embodiment uses three effective characteristic parameters of chicken carcass A length, body width and projected area to predict the actual weight of chicken carcass A and output the grading result. The accuracy rate of judging "too small" chicken carcass A reaches 94.3%, and the accuracy rate of judging "too large" chicken carcass A reaches 95.6%.

[0042] Example 5 This embodiment provides a method for intelligent grading and segmenting of chicken carcasses based on machine vision. The method is applied to the intelligent grading and segmenting system for chicken carcasses based on machine vision in the above-mentioned embodiment 1. When the method is applied to segmenting a chicken carcass A into multiple parts, the segmentation station includes an automated segmentation system. The method includes: Step 1: Hang the feet of the chicken carcass A on the hook 12, let the chicken carcass A hang naturally, and face the camera 21 with the chicken breast side, and the chicken carcass A moves along the slide rail 11 with the hook 12; Step 2: After the fiber optic sensor of the first collection station 2 detects the chicken carcass A, it triggers the background light board 22 to flash and the camera 21 to take pictures, and uploads the image information to the industrial computer; The chicken carcass A enters the second collection station 3. After the optical fiber sensor of the second collection station 3 detects the chicken carcass A, it triggers the background light board 22 to flash and the camera 21 to take pictures, and uploads the image information to the industrial computer; Step 3: According to the grading results obtained in step 3, the industrial computer analyzes and processes the image information, and automatically extracts three effective characteristic parameters of the chicken carcass A length, body width and projection area; The body width is selected as a single characteristic parameter input, and according to the pre-constructed regression model for length grade prediction: W1=0.132W+2.131 (R²=0.999, P<0.01), (W1 is the predicted value of the actual body width of the chicken carcass A, W is the characteristic value of the body width of the chicken carcass A, and the coefficient of this model is as high as 0.999, which can well predict the actual width of the chicken carcass), the predicted value of the actual width of the chicken carcass is calculated, and the predicted value of the actual width of the chicken carcass is used as the output, and the body width grade is automatically determined by comparing the predicted value of the body width with the preset value of the chicken carcass grading width, and the obtained grading result is displayed on the display screen 6; Selecting three effective characteristic parameters of the length, body width and projection area of ​​the chicken carcass A as input, according to the pre-built regression model for weight grade prediction: M=0.423W+0.001S+0.017L-901.644 (R²=0.968, P<0.01), (M is the predicted value of the actual weight of the chicken carcass A, W is the characteristic value of the body width of the chicken carcass A, S is the characteristic value of the projection area of ​​the chicken carcass A, L is the characteristic value of the length of the chicken carcass A, the coefficient of the model is as high as 0.968, which can well predict the actual weight of the chicken carcass A), calculate the predicted value of the actual weight of the chicken carcass A, and use the predicted value of the weight of the chicken carcass A as output, and automatically determine the weight grade of the chicken carcass A by comparing the predicted value of the weight of the chicken carcass A with the preset value of the grading weight of the chicken carcass A, and display the obtained grading result on the display screen 6; Step 4: The ejection cylinder 41 sequentially ejects the chicken carcasses A that are not suitable for the width and weight of the automated cutting system; Step 5: The remaining chicken carcass A can smoothly enter the process-based automated cutting system to automatically cut the chicken carcass A into multiple parts such as the whole leg, three-section wing, and chicken breast.

[0043] It should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A chicken carcass intelligent grading and segmentation system based on machine vision, characterized in that: It includes a raw material conveying device, an industrial computer, and a collection station, a rejection station and a segmentation station sequentially arranged along the raw material conveying direction; The collection station includes a camera and a background light board respectively located on both sides of the raw material conveying device. The chicken carcass is hung upside down on the raw material conveying device with the chest of the chicken carcass facing the camera. The camera collects the image information of the chicken carcass and uploads it to the industrial computer. The industrial computer grades the specifications of the chicken carcass according to the image information. The rejecting station includes a rejecting device and a material discharge chute respectively located on both sides of the raw material conveying device. A material discharge conveying device is arranged below the material discharge chute. The industrial computer controls the rejecting device to reject the chicken carcasses that do not meet the requirements according to the grading results of the chicken carcasses, and the chicken carcasses that meet the requirements continue to enter the cutting station with the raw material conveying device for cutting.

2. The machine vision-based intelligent chicken carcass grading and segmentation system according to claim 1, characterized in that: The collecting station, the rejecting station and the dividing station are all provided with incoming material sensing devices.

3. The machine vision-based chicken carcass intelligent grading and segmentation system according to claim 1, characterized in that: The industrial computer is connected with a display screen and a human-computer interaction touch panel.

4. The machine vision-based chicken carcass intelligent grading and segmentation system according to claim 1, characterized in that: The number of the collection stations is 2, which are a first collection station and a second collection station which are arranged in sequence.

5. The machine vision-based chicken carcass intelligent grading and segmentation system according to claim 1, characterized in that: The specifications of the chicken carcass include the length, body width and weight of the chicken carcass; The industrial computer obtains the length, body width and projection area of ​​the chicken carcass as characteristic parameters through the image information collected by the camera, inputs the characteristic parameters as a single or combined input, establishes a regression model for predicting the grade of the chicken carcass, and automatically determines the different specification grades of the chicken carcass.

6. The machine vision-based chicken carcass intelligent grading and segmentation system according to claim 5, characterized in that: When the chicken carcass is headed, the segmentation station includes an automated head-cutting device, and the industrial computer inputs the length of the chicken carcass as a single characteristic parameter, establishes a regression model for predicting the length grade of the chicken carcass, and automatically determines the length grade of the chicken carcass; When the chicken carcass is wing-cut, the cutting station includes an automated wing-cutting device, and the industrial computer inputs the body width of the chicken carcass as a single characteristic parameter, establishes a regression model for predicting the body width grade of the chicken carcass, and automatically determines the body width grade of the chicken carcass; When the chicken carcass is cut into multiple parts, the cutting station is an automated cutting system, and the industrial computer inputs the length, body width and projection area of ​​the chicken carcass as combined characteristic parameters, establishes a regression model for predicting the weight grade of the chicken carcass, and automatically determines the weight grade of the chicken carcass.

7. The machine vision-based chicken carcass intelligent grading and segmentation system according to claim 6, characterized in that: When the chicken carcass is divided into multiple parts, the industrial computer inputs the body width of the chicken carcass as a single characteristic parameter, establishes a regression model for predicting the body width grade of the chicken carcass, and automatically determines the body width grade of the chicken carcass; The industrial computer uses the length, body width and projection area of ​​the chicken carcass as combined characteristic parameters, establishes a regression model for predicting the weight grade of the chicken carcass, and automatically determines the weight grade of the chicken carcass; The chicken carcasses that meet the requirements are selected based on the weight and body width of the chicken carcasses and enter the cutting station for cutting into multiple parts.

8. A method for intelligent grading and segmenting of chicken carcasses based on machine vision, the method being applied to the intelligent grading and segmenting system for chicken carcasses based on machine vision according to any one of claims 1 to 7, characterized in that: The method comprises: Step 1: Hang the chicken carcass naturally with the breast side facing the camera; Step 2: The chicken carcass is moved to the collection station to collect images of the chicken carcass, and the image information is uploaded to the industrial computer; Step 3: the industrial computer analyzes and processes the image information, selects required feature parameters, establishes a regression model for predicting the chicken carcass grade, and automatically determines the chicken carcass grade; Step 4: According to the grade results, the chicken carcasses that do not meet the requirements are removed, and the remaining chicken carcasses continue to move to the cutting station; Step 5: Cut the chicken carcass into corresponding parts at the cutting station.

9. The method for intelligent grading and segmenting of chicken carcasses based on machine vision according to claim 8, characterized in that: When beheading a chicken carcass, In step 3, the length of the chicken carcass is selected as a single characteristic parameter input, a regression model for predicting the length grade of the chicken carcass is established, and the length grade of the chicken carcass is automatically determined; In step 5, the segmentation station is an automated head cutting device, and the head of the chicken carcass is segmented by the automated head cutting device; When wing cutting is performed on chicken carcasses, In step 3, the body width of the chicken carcass is selected as a single characteristic parameter input, a regression model for predicting the body width grade of the chicken carcass is established, and the body width grade is automatically determined; In step 5, the segmentation station is an automated wing cutting device, and the wings of the chicken carcass are segmented by the automated wing cutting device; When cutting a chicken carcass into multiple parts, In step 3, the length, body width and projected area of ​​the chicken carcass are selected as combined characteristic parameter inputs, a regression model for predicting the weight grade of the chicken carcass is established, and the weight grade is automatically determined; In step 5, the cutting station is an automated cutting system, through which the wings, legs and breasts of the chicken carcass are cut.

10. The method for intelligent grading and segmenting of chicken carcasses based on machine vision according to claim 9, characterized in that: The regression model for length grade prediction is L1=0.211L-4.521, wherein L1 is the predicted value of the actual length of the chicken carcass, and L is the characteristic value of the chicken carcass length; The regression model for predicting the body width grade is W1=0.132W+2.131, wherein W1 is the predicted value of the actual body width of the chicken carcass, and W is the characteristic value of the body width of the chicken carcass; The regression model for weight grade prediction is M=0.423W+0.001S+0.017L-901.644, wherein M is the predicted value of the actual weight of the chicken carcass, W is the characteristic value of the width of the chicken carcass, S is the characteristic value of the projected area of ​​the chicken carcass, and L is the characteristic value of the length of the chicken carcass.

Images