CN116912887B - A broiler breeding management method and system - Google Patents

Abstract

The invention belongs to the technical field of data collection and intelligent breeding. It proposes a broiler breeding management method and system. Specifically, it collects broiler data through an industrial CCD camera to obtain an original picture of the broiler, and then preprocesses the original picture of the broiler to form a detection image, and then build a texture pre-approval model by detecting the image to obtain the pre-approval reference value, and finally build a database for broiler breeding management based on the pre-approval reference value. During the image data collection process, pre-approved reference values are used to improve the accuracy of identifying tendons, membranes or fat in broiler images, and form graphical quantitative data for quantification of morphological characteristics of areas of interest in images, which greatly improves the understanding of broiler populations or broiler breeding. The accuracy of the model built in it effectively provides better data support for model judgment and quantitative results, thereby improving the accuracy of applying graph data to big data models.

Description

A broiler breeding management method and system

Technical field

The invention belongs to the technical fields of data collection and intelligent breeding, and specifically relates to a broiler breeding management method and system.

Background technique

Broiler chickens play an important role in agriculture and food production. Their meat is widely used in food processing, catering industry and household consumption. In modern broiler chicken breeding, broiler farmers use genetic selection and optimization through broiler breeding management. Cultivate broilers that grow quickly and have high breast muscle content. During the cultivation of broiler chickens, the fat accumulation of chicken breasts is one of the key quality factors of broilers. The traditional cultivation direction is to sample broiler chickens through chicken slices and quantify the fat accumulation of broiler chickens, thereby forming the quality of broiler chickens to further improve the quality of broiler chickens. Optimizing broiler breeding management.

However, since this method is based on images of chicken slice processing, the processing method is usually complex and cumbersome. Improper handling of the sampled broilers can even easily lead to a waste of resources or an increase in operating costs. Therefore, the existing technology can generally use broilers to perform broiler processing in production or assembly lines. Image capture: collect images of broilers that have been cleaned after all internal organs have been removed, and then use big data to quantify the quality of different batches of broilers. However, when collecting image data of broiler chickens in the assembly line, there is often insufficient accuracy in judging tendons, membranes or fat in the broiler chicken images, which in turn affects the evaluation results or the judgment accuracy of the model. Therefore, in order to further improve the accuracy and stability of broiler chicken breeding management sex, it is urgent to optimize the obtained broiler images.

Contents of the invention

The purpose of the present invention is to propose a broiler chicken breeding management method and system to solve one or more technical problems existing in the prior art and at least provide a beneficial choice or creation condition.

In order to achieve the above objects, according to one aspect of the present invention, a broiler chicken breeding management method is provided, which method includes the following steps:

S100, collects broiler data through industrial CCD cameras to obtain original images of broilers;

S200, preprocess the original broiler image to form a detection image;

S300, construct a texture pre-approval model by detecting images and obtain a pre-approval reference value;

S400: Construct a broiler breeding management database based on pre-approved reference values.

Further, in step S100, the method of collecting broiler data through an industrial CCD camera to obtain the original picture of the broiler is: the CCD camera is an area array CCD camera or a linear array CCD camera; in the broiler production line, all internal organs are removed And the image of the cleaned broiler is collected, and the collected image of the broiler is used as the original picture of the broiler.

Further, in step S200, after preprocessing the original image of the broiler chicken, the method of forming the detection image is: performing grayscale processing on the original image of the broiler chicken, and using edge detection based on the Canny operator, the Sobel operator or the Laplacian operator. The algorithm intercepts the area of interest from the original image of the broiler chicken, performs image erosion on the intercepted image, and uses the final image as the detection image.

Further, in step S300, the texture pre-batch model is constructed by detecting the image, and the method of obtaining the pre-batch reference value is: binarizing the detection image. The algorithm used in the binarization process is the OTSU method, and the obtained The image is recorded as the detection binary image; the detection binary image is divided into regions, and the areas with a pixel value of 255 in the detection binary image are respectively used as information identification areas WG.

Further, in step S300, the texture pre-batch model is constructed by detecting the image, and the method of obtaining the pre-batch reference value is: defining the total number of pixels in the information identification area WG as the patch area metric value HPLt, and obtaining each information identification area. The patch metric value HPLt in the discrimination area WG forms a sequence as the first patch sequence HPLt_Ls. The arithmetic mean of all elements in the first patch sequence is calculated and called the screening limit amount LmtLs; the small number in the first patch sequence is calculated A new sequence composed of each element equal to the screening limit amount is recorded as the second patch sequence R_HPLt_Ls; the arithmetic mean of all elements in the second patch sequence is recorded as the bounded mean E_LmtLs; the pre-batch reference value is recorded as FRV, The calculation method is:

;

Among them, exp() is an exponential function with the natural constant e as the base, ds<> is a range function, and the result of the range function is the difference between the maximum value and the minimum value in the calling sequence.

Since the pre-batch reference value is calculated based on the binary image of the detection image, it effectively quantifies the morphological characteristics of the area of interest in the image to form graphical quantification data. However, in the case of insufficient brightness in the image collection environment, the above method is used The calculated pre-approval reference value FRV may be insufficient in quantification. This is because this method reduces the sensitivity of the relationship between the contour shape of each information identification area and the corresponding patch area measurement value, resulting in the ideal filtered results. The accuracy of the patch-area metric values is lost, and there is currently no feasible technology to enhance the sensitivity of the relationship between the contour form of each information identification area WG and the corresponding patch-area metric value. In order to eliminate the outline form of the information identification area WG Regarding the impact on the slice domain metric value it has, the present invention proposes a more preferred solution.

Preferably, in step S300, a texture pre-batch model is constructed by detecting images, and the method of obtaining the pre-batch reference value is:

If there is at least one pixel with a grayscale value of 0 in the eight neighborhoods of a certain pixel in the information identification area WG, then this pixel is defined as the boundary pixel of the information identification area WG; in an information identification area WG Within , take any pixel to the nearest boundary pixel to the pixel as the near boundary pixel of the pixel, and take the distance between a pixel and its corresponding near boundary pixel as the boundary of the pixel Distance EDis, obtain the EDis of all pixels in the information identification area WG to form a sequence as a path sequence;

The maximum value in the path sequence is used as the anisotropic path EDL of the information identification area WG; the anisotropic path of each information identification area WG is obtained to form a sequence as an anisotropic path sequence EDL_Ls; the information identification is obtained by calculation The average amplitude deviation of area WG is from GRAD. Taking n as the serial number of the information identification area, the calculation formula of the average amplitude deviation from GRAD _n of the nth information identification area is as follows:

;

where mean<> is the arithmetic mean function, EDL _n is the calculated anisotropic path EDL of the nth information identification area in the detection binary image; calculate the screening domain GRMD of the detection image:

;

Among them, BecEDL<> is the standard deviation function of the call sequence obtained after Bessel correction. cpf represents the statistical compensation coefficient. Its value range is cpf∈[1.05,1.15], and its default value is set to 1.10; if n information identification areas meet the condition GRADn ≤ GRMD, then the information identification area is defined as the first information identification area RWG, and the opposite direction connection path of the first information identification area is defined as the first opposite direction connection path REDL; each The first anisotropic connection path REDL constitutes the first anisotropic connection path sequence REDL_Ls; the pre-approved reference value is recorded as FRV, and the calculation method is as follows:

;

Among them, i1 is the accumulated variable, k is the number of the first information identification area, and exp() is the exponential function with the natural constant e as the base.

Beneficial effects: Since the most significant heterodirectional paths in the information identification area WG are used to screen the information identification area WG, the influence of the contour shape of the information identification area WG on the screening is effectively eliminated; at the same time, a statistical compensation coefficient is introduced As a secondary screening, it is ensured that the first information identification area with higher quantitative advantage can be obtained, thereby improving the accuracy of the obtained quantitative model of the first information identification area and the sensitivity of the associated attributes. During the image data collection process, pre-approved reference values are used to improve the accuracy of identifying tendons, membranes or fat in broiler images, and form graphical quantitative data for quantification of morphological characteristics of areas of interest in images, which greatly improves the understanding of broiler populations or broiler breeding. The accuracy of the model built in it effectively provides better data support for model judgment and quantitative results, thereby improving the accuracy of applying graph data to big data models.

Further, in step S400, the method of constructing a database for broiler breeding management based on the pre-batch reference values is: using each detection image obtained from the broiler chickens of the same production batch as the same batch detection chart sequence, when the collection of the same batch detection chart sequence is completed , or the broilers of the same production batch have obtained detection images, then obtain the pre-approval reference values of the detection images respectively, and construct a sequence for each obtained pre-approval reference value as the pre-batch sequence; use the pre-batch sequence to construct a box plot , and use box plots to judge abnormal data to obtain several outliers, remove the detection images corresponding to each outlier from the same batch of detection map sequences, and then store the same batch of detection map sequences into a database used for broiler breeding management.

Preferably, all undefined variables in the present invention, if not clearly defined, can be manually set thresholds.

The present invention also provides a broiler chicken breeding management system, which includes: a processor, a memory and a computer program stored in the memory and executable on the processor. The processor When the computer program is executed, the steps in the broiler chicken breeding management method are implemented. The broiler chicken breeding management system can be run on computing devices such as desktop computers, notebook computers, handheld computers, and cloud data centers. The running system may include, but is not limited to, a processor, a memory, and a server cluster. The processor executes the computer program and runs in the following system units:

Image acquisition unit, used to collect broiler data through industrial CCD cameras to obtain original images of broilers;

The image preprocessing unit is used to preprocess the original broiler image to form a detection image;

A model building unit, used to build a texture pre-batch model by detecting images and obtain a pre-batch reference value;

A database building unit used to build a database for broiler breeding management based on pre-approved reference values;

The beneficial effects of the present invention are: during the image data acquisition process through pre-approved reference values, the accuracy of identifying tendons, membranes or fat in broiler images is improved, and graphic quantified data is formed for quantification of morphological characteristics of areas of interest in the image, which greatly Improve the accuracy of models built in broiler populations or broiler breeding, effectively provide better data support for model judgment and quantitative results, thereby improving the accuracy of graphical data applied to big data models, by comparing fat types in broiler images The interference components with similar pixels are quantified, and the quantified values are checked for abnormalities, and the redundant data or abnormal images in the collected broiler images are efficiently and accurately screened out, thereby improving the obtained features for extracting the fat content of broilers. Learning data or training data provides more accurate and reliable data support for broiler breeding strategy adjustment and management. With the help of this technology, because farms can more accurately assess the fat content of broilers, potential problems, such as unbalanced broiler feeds or genetic diseases, can be quickly discovered by analyzing abnormalities in quantitative values. More accurate data reserves can help farms take correct measures. Accurate database reserves provide researchers and farm operators in the field of broiler breeding with more accurate information, and can also better optimize the breeding environment and management strategies.

Description of drawings

The above and other features of the present invention will be more apparent from the detailed description of the embodiments shown in the accompanying drawings. In the drawings of the present invention, the same reference numerals designate the same or similar elements. It will be apparent that the appended drawings in the following description The drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without exerting creative efforts. In the drawings:

Figure 1 shows a flow chart of a broiler chicken breeding and management method;

Figure 2 shows the structure diagram of a broiler breeding management system.

Detailed ways

The following will give a clear and complete description of the concept, specific structure and technical effects of the present invention in conjunction with the embodiments and drawings, so as to fully understand the purpose, solutions and effects of the present invention. It should be noted that, as long as there is no conflict, the embodiments and features in the embodiments of this application can be combined with each other.

Figure 1 shows a flow chart of a broiler chicken breeding and management method. The following is a description of a broiler chicken breeding and management method according to an embodiment of the present invention in conjunction with Figure 1. The method includes the following steps:

S100, collects broiler data through industrial CCD cameras to obtain original images of broilers;

S200, preprocess the original broiler image to form a detection image;

S300, construct a texture pre-approval model by detecting images and obtain a pre-approval reference value;

S400: Construct a broiler breeding management database based on pre-approved reference values.

Further, in step S100, the method of collecting broiler data through an industrial CCD camera to obtain the original picture of the broiler is: the CCD camera is an area array CCD camera or a linear array CCD camera; in the broiler production line, all internal organs are removed And the image of the cleaned broiler is collected, and the collected image of the broiler is used as the original picture of the broiler.

Further, in step S200, after preprocessing the original image of the broiler chicken, the method of forming the detection image is: performing grayscale processing on the original image of the broiler chicken, and using edge detection based on the Canny operator, the Sobel operator or the Laplacian operator. The algorithm intercepts the area of interest from the original image of the broiler chicken, performs image erosion on the intercepted image, and uses the final image as the detection image.

Further, in step S300, the texture pre-batch model is constructed by detecting the image, and the method of obtaining the pre-batch reference value is: binarizing the detection image. The algorithm used in the binarization process is the OTSU method, and the obtained The image is recorded as the detection binary image; the detection binary image is divided into regions, and the areas with a pixel value of 255 in the detection binary image are respectively used as information identification areas WG.

Further, in step S300, the texture pre-batch model is constructed by detecting the image, and the method of obtaining the pre-batch reference value is: defining the total number of pixels in the information identification area WG as the patch area metric value HPLt, and obtaining each information identification area. The patch metric value HPLt in the discrimination area WG forms a sequence as the first patch sequence HPLt_Ls. The arithmetic mean of all elements in the first patch sequence is calculated and called the screening limit amount LmtLs; the small number in the first patch sequence is calculated A new sequence composed of each element equal to the screening limit amount is recorded as the second patch sequence R_HPLt_Ls; the arithmetic mean of all elements in the second patch sequence is recorded as the bounded mean E_LmtLs; the pre-batch reference value is recorded as FRV, The calculation method is:

;

Among them, exp() is an exponential function with the natural constant e as the base, ds<> is a range function, and the result of the range function is the difference between the maximum value and the minimum value in the calling sequence.

Preferably, in step S300, a texture pre-batch model is constructed by detecting images, and the method of obtaining the pre-batch reference value is:

If there is at least one pixel with a grayscale value of 0 in the eight neighborhoods of a certain pixel in the information identification area WG, then this pixel is defined as the boundary pixel of the information identification area WG; in an information identification area WG Within , take any pixel to the nearest boundary pixel to the pixel as the near boundary pixel of the pixel, and take the distance between a pixel and its corresponding near boundary pixel as the boundary of the pixel Distance EDis, obtain the EDis of all pixels in the information identification area WG to form a sequence as a path sequence;

The maximum value in the path sequence is used as the anisotropic path EDL of the information identification area WG; the anisotropic path of each information identification area WG is obtained to form a sequence as an anisotropic path sequence EDL_Ls; the information identification is obtained by calculation The average amplitude deviation of area WG is from GRAD. Taking n as the serial number of the information identification area, the calculation formula of the average amplitude deviation from GRAD _n of the nth information identification area is as follows:

;

where mean<> is the arithmetic mean function, EDL _n is the calculated anisotropic path EDL of the nth information identification area in the detection binary image; calculate the screening domain GRMD of the detection image:

;

Among them, BecEDL<> is the standard deviation function of the call sequence obtained after Bessel correction, and cpf represents the statistical compensation coefficient. Its value range is cpf∈[1.05,1.15], and its default value is set to 1.10;

If the nth information identification area meets the condition GRADn ≤ GRMD, then the information identification area is defined as the first information identification area RWG, and the opposite direction connection path of the first information identification area is defined as the first opposite direction connection path REDL ;Each first anisotropic connection REDL constitutes a first anisotropic connection sequence REDL_Ls; The pre-approval reference value is recorded as FRV, and the calculation method is as follows:

;

Among them, i1 is the accumulated variable, k is the number of the first information identification area, and exp() is the exponential function with the natural constant e as the base.

Further, in step S400, the method of constructing a database for broiler breeding management based on the pre-batch reference values is: using each detection image obtained from the broiler chickens of the same production batch as the same batch detection chart sequence, when the collection of the same batch detection chart sequence is completed , or the broilers of the same production batch have obtained detection images, then obtain the pre-approval reference values of the detection images respectively, and construct a sequence for each obtained pre-approval reference value as the pre-batch sequence; use the pre-batch sequence to construct a box plot , and use box plots to judge abnormal data to obtain several outliers, remove the detection images corresponding to each outlier from the same batch of detection map sequences, and then store the same batch of detection map sequences into a database used for broiler breeding management.

A broiler chicken breeding management system provided by an embodiment of the present invention. Figure 2 is a structural diagram of a broiler chicken breeding management system of the present invention. The broiler chicken breeding management system of this embodiment includes: a processor, a memory and a storage device. A computer program is stored in the memory and can be run on the processor. When the processor executes the computer program, the steps in the above embodiment of the broiler breeding management system are implemented.

The system includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor executes the computer program to run in a unit of the following system:

Image acquisition unit, used to collect broiler data through industrial CCD cameras to obtain original images of broilers;

The image preprocessing unit is used to preprocess the original broiler image to form a detection image;

A model building unit, used to build a texture pre-batch model by detecting images and obtain a pre-batch reference value;

A database building unit used to build a database for broiler breeding management based on pre-approved reference values;

The broiler breeding management system can be run on computing devices such as desktop computers, notebook computers, handheld computers, and cloud servers. The executable system of the broiler chicken breeding management system may include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the examples described are only examples of a broiler chicken breeding management system, and do not constitute a limitation to a broiler chicken breeding management system. It may include more or fewer components than the examples, or some combinations. Components, or different components, for example, the broiler breeding management system may also include input and output devices, network access devices, buses, etc.

The so-called processor can be a central processing unit (Central Processing Unit, CPU), or other general-purpose processor, digital signal processor (Digital Signal Processor, DSP), application specific integrated circuit (Application Specific Integrated Circuit, ASIC), on-site Programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general processor can be a microprocessor or the processor can be any conventional processor, etc. The processor is the control center of the operating system of the broiler breeding management system and uses various interfaces and lines to connect the entire system. The broiler chicken breeding management system can run various parts of the system.

The memory may be used to store the computer program and/or module, and the processor implements the process by running or executing the computer program and/or module stored in the memory and calling data stored in the memory. Various functions of broiler chicken breeding management system. The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the storage data area may store Data created based on the use of mobile phones (such as audio data, phone books, etc.), etc. In addition, the memory can include high-speed random access memory, and can also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart memory card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card , Flash Card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

While the invention has been described in considerable detail and particularly with respect to several of the described embodiments, it is not intended to be limited to any such details or embodiments or to any particular embodiment so as to effectively encompass the intended scope of the invention. In addition, the above description of the present invention is based on embodiments foreseeable by the inventor for the purpose of providing a useful description, and those non-substantive changes to the present invention that are not yet foreseen can still represent equivalent changes of the present invention.

Claims (5)

1. A broiler breeding management method, characterized in that the method comprises the following steps:

s100, acquiring data of the broiler chickens through an industrial CCD camera to obtain an original image of the broiler chickens;

s200, preprocessing an original image of the broiler chicken to form a detection image;

s300, constructing a texture pre-batch model by detecting an image to obtain a pre-batch reference value;

s400, constructing a database for broiler breeding management according to the pre-batch reference values;

in step S300, a texture pre-batch model is constructed by detecting an image, and the method for obtaining the pre-batch reference value is as follows: performing binarization processing on the detected image, wherein an algorithm adopted in the binarization processing is an OTSU method, and the obtained image is recorded as a detected binary image; dividing the detection binary image into areas, and taking the areas with the pixel values of 255 in the detection binary image as information identification areas WG;

if at least one gray value of a pixel point exists in eight adjacent areas of a certain pixel point in the information identification area WG, defining the pixel point as a boundary area pixel of the information identification area WG; in an information authentication zone WG; for any pixel point, obtaining the distance between each boundary region pixel and the pixel point, and taking the boundary region pixel with the smallest distance as a near-boundary region pixel; taking the distance between a pixel point and a corresponding near-field pixel as the field distance EDis of the pixel point, and acquiring the field distances of all the pixel points in the information identification area WG to form a sequence as a path sequence;

taking the maximum value in the path sequence as the information discrimination area WGDifferent-direction diameter EDL; the different paths of the information authentication areas WG are obtained to form a sequence which is used as a different path sequence EDL_Ls; calculating the average deviation GRAD of the information identification area WG, taking n as the serial number of the information identification area, and then the average deviation GRAD of the nth information identification area _n The calculation formula of (2) is as follows:

；

wherein mean<>EDL as a function of arithmetic mean _n Is the calculated different-direction diameter EDL of the nth information identification area in the detection binary diagram; calculating a screening bound domain GRMD of the detected image:

；

wherein BecEDL<>For the standard deviation function of the call sequence obtained after Bessel correction, cpf represents a statistical compensation coefficient with a value range of cpf E [1.05,1.15]]Setting the default value to be 1.10; if the nth information authentication zone meets the GRAD _n Defining the information identification area as a first information identification area RWG and defining the different-direction access of the first information identification area as a first different-direction access REDL if GRMD is not more than; each first differential path REDL forms a first differential path sequence redl_ls; the pre-batch reference value is denoted as FRV and is calculated as follows:

；

wherein i1 is an accumulation variable, k is the number of first information identification areas, exp () is an exponential function with a natural constant e as a base.

2. The method for managing broiler breeding according to claim 1, wherein in step S100, the method for acquiring the broiler data by using an industrial CCD camera to obtain the original image of broiler is as follows: the CCD camera is an area array CCD camera or a linear array CCD camera; in a production line of broiler chicken production, image acquisition is carried out on broiler chicken with all viscera removed and cleaned, and an acquired image of the broiler chicken is used as an original image of the broiler chicken.

3. The method for managing broiler breeding according to claim 1, wherein in step S200, after preprocessing the original image of broiler chicken, the method for forming the detection image is as follows: gray processing is carried out on the original graph of the broiler chicken, a region of interest is cut out from the original graph of the broiler chicken through an edge detection algorithm based on a Canny operator, a Sobel operator or a Laplacian operator, image corrosion is carried out on the cut-out image, and the finally obtained image is used as a detection image.

4. The method according to claim 1, wherein in step S400, the method for constructing a database for broiler breeding management from the pre-batch reference values is as follows: taking all detection images obtained by broilers in the same production batch as a same batch detection image sequence, respectively obtaining the pre-batch reference values of the detection images when the collection of the same batch detection image sequence is completed or the detection images are obtained by broilers in the same production batch, and constructing a sequence by using all the obtained pre-batch reference values as a pre-batch sequence; and constructing a box-type diagram by using the pre-batch sequence, judging abnormal data through the box-type diagram to obtain a plurality of abnormal values, removing detection images corresponding to the abnormal values from the same batch of detection diagram sequences, and storing the same batch of detection diagram sequences into a database applied to broiler breeding management.

5. A broiler breeding management system, characterized in that it comprises: a processor, a memory and a computer program stored in the memory and executable on the processor, the processor implementing the steps of a broiler breeding management method according to any of claims 1-4 when the computer program is executed, the broiler breeding management system running on a desktop computer, a notebook computer, a palm computer or a cloud server.