An egg laying performance evaluation device, method, system and electronic equipment

Abstract

The application discloses an egg-laying hen production performance evaluation device, method, system and electronic equipment, and relates to the field of farms management.The method comprises the following steps: acquiring an egg image under infrared light and infrared laser irradiation collected by an infrared camera, and acquiring egg quantity and egg quality information according to the egg image; acquiring a feed image under infrared laser irradiation collected by the infrared camera, calculating a feed remaining volume according to the feed image, and further obtaining a feed intake of a chicken; and generating an egg-laying hen production performance distribution map based on an egg-laying hen production performance evaluation model according to the egg quantity, the egg quality information and the feed intake of the chicken. The method has a high accuracy in evaluating the production performance of the egg-laying hen, can timely locate a position of an abnormal chicken in production performance, and can obtain a relatively accurate egg-laying hen production performance distribution map, so that the distribution of the production performance of the chicken in an egg-laying hen or a breeding hen house can be intuitively and clearly understood.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of chicken farm management, and in particular to an egg laying performance evaluation device, method, system and electronic equipment. BACKGROUND

[0002] With the transformation and upgrading of modern animal husbandry, new requirements are put forward for the feeding and management of cage livestock and poultry. In the process of breeding of laying hens and breeding hens, abnormal eggs are inevitable. Abnormal eggshells are generally thin and can easily break during normal collection, which can contaminate the egg belt. Therefore, abnormal eggs need to be picked up from the egg belt in time to avoid affecting other eggs.

[0003] In addition, China is the world's largest egg producer and consumer. In the process of breeding laying hens, the production performance of laying hens is a key indicator of chicken farm management and a direct determinant of breeding efficiency. In order to ensure the economic benefits of egg production, it is crucial to accurately understand the production performance of laying hens in a timely manner.

[0004] At present, the production performance of individual chickens is completely dependent on manual qualitative detection. The stacked cage breeding mode has high density, many layers, high height, dim and large changes in light, which makes the manual inspection method have problems such as high labor intensity, incomplete inspection, high subjectivity, and frequent contact between people and chickens. Therefore, the method of manual qualitative detection for evaluating the production performance of laying hens has the defects of low evaluation efficiency, poor accuracy and strong subjectivity in the evaluation process. SUMMARY

[0005] The purpose of the present application is to provide an egg laying performance evaluation device, method, system and electronic equipment to improve the efficiency and accuracy of egg laying performance evaluation.

[0006] To achieve the above-mentioned purpose, the present application provides the following solutions:

[0007] An egg laying performance evaluation device, comprising: a host computer and a chicken coop inspection module;

[0008] The chicken coop inspection module comprises n image acquisition units and a positioning unit; each layer of chicken coop is provided with an image acquisition unit; the image acquisition unit comprises a first camera and a second camera; the first camera is used for acquiring a feed image; the first camera is provided with a linear laser; the second camera is used for acquiring an egg image; the second camera is provided with a linear laser and a light source; the positioning unit is used for acquiring the position of the chicken coop;

[0009] The host computer is connected with the image acquisition unit and the positioning unit respectively; the host computer is used for determining the number of eggs and egg quality information according to the egg image, determining the feed intake of the chicken according to the feed image, and determining the production performance of the laying hen according to the number of eggs, the egg quality information and the feed intake of the chicken; the egg quality information includes eggshell quality information, egg geometric shape parameters and egg weight; the eggshell quality information includes normal eggs, soft-shelled eggs and thin-shelled eggs.

[0010] Optionally, the first camera and the second camera are both near-infrared cameras; the light of the line laser and the light source is near-infrared light; the near-infrared camera adopts a band-pass filter; the imaging wavelength of the near-infrared camera is consistent with the wavelength of the line laser and the light source.

[0011] Optionally, the henhouse inspection module further comprises a moving unit, a first support, a second support and a fixing block.

[0012] One end of the first support is fixed to the moving unit; the other end of the first support is provided with the positioning unit; one end of the second support is connected with the first support through the fixing block; the other end of the second support is provided with the image acquisition unit.

[0013] Optionally, the image acquisition unit comprises a first single-hole expansion seat, a second single-hole expansion seat, a first camera angle adjustment subunit and a second camera angle adjustment subunit.

[0014] The first camera angle adjustment subunit is connected with the first support through the first single-hole expansion seat; the second camera angle adjustment subunit is connected with the second support through the second single-hole expansion seat; the first camera is connected with the first camera angle adjustment subunit; the second camera is connected with the second camera angle adjustment subunit.

[0015] A method for evaluating the production performance of laying hens, which is applied to the device for evaluating the production performance of laying hens, and comprises the following steps:

[0016] An egg image, a feed image and corresponding cage position information are acquired; the egg image is acquired by a second camera; the feed image is acquired by a first camera; the cage position information is acquired by a positioning unit;

[0017] According to the egg image, an egg detection model is used to determine egg quantity and egg quality information; the egg quality information includes eggshell quality information, egg geometric shape parameters and egg weight; the eggshell quality information includes normal eggs, soft-shelled eggs and thin-shelled eggs; wherein the egg detection model includes an egg segmentation model and an egg evaluation model; the egg segmentation model is obtained by training a segmentation network using a first training data set; the first training data set includes a plurality of historical egg images with labeled egg pixels and laser line pixels; the egg evaluation model is obtained by training a support vector machine using a second training data set; the second training data set includes egg pixels, laser line pixels of historical egg images and corresponding egg quantity and egg quality information.

[0018] According to the feed image and the corresponding chicken coop position information, the amount of feed eaten by the chicken is determined; the amount of feed eaten by the chicken is the weight of the feed eaten by the chicken;

[0019] According to the egg quantity, the egg quality information and the amount of feed eaten by the chicken, an egg-laying performance evaluation model is used to determine an egg-laying performance score; wherein the egg-laying performance evaluation model is obtained by training a machine learning model using a third training data set; the third training data set includes an input data set and an output data set; the input data set includes the egg quantity, the egg quality information and the amount of feed eaten by the chicken of the training egg-laying hen; the output data set is the production performance score of the training egg-laying hen by an expert.

[0020] Optionally, it further includes:

[0021] According to the egg quantity, the egg quality information, the amount of feed eaten by the chicken, the chicken coop position information and the egg-laying performance score, an egg position distribution map, an egg quality information distribution map, a chicken feed intake distribution map and an egg-laying performance distribution map are generated.

[0022] Optionally, according to the egg image, an egg detection model is used to determine egg quantity and egg quality information, specifically including:

[0023] According to the egg image, the egg segmentation model is used to segment the egg pixels and the laser line pixels;

[0024] According to the laser line pixels and the camera imaging principle, the three-dimensional point cloud of the egg is determined;

[0025] According to the egg pixels, the egg contour and the egg texture information are determined;

[0026] According to the egg contour, the egg texture information and the three-dimensional point cloud of the egg, the egg evaluation model is used to determine the egg quantity and the egg quality information.

[0027] Optionally, according to the feed image and corresponding chicken coop position information, the chicken feed intake is determined, specifically comprising:

[0028] According to the feed image and corresponding chicken coop position information, the surface three-dimensional point cloud data of the remaining feed in the trough is determined;

[0029] According to the surface three-dimensional point cloud data, in combination with the trough size and the initial volume of the feed in the trough, the chicken feed intake volume is determined;

[0030] According to the chicken feed intake volume and the feed density, the chicken feed intake is determined.

[0031] An egg laying hen production performance evaluation system, comprising:

[0032] An image acquisition module for acquiring egg images, feed images and corresponding chicken coop position information;

[0033] An egg detection module for determining the number of eggs and egg quality information according to the egg images by using an egg detection model; the egg quality information includes normal eggs, soft-shelled eggs and thin-shelled eggs; wherein the egg detection model comprises an egg segmentation model and an egg evaluation model; the egg segmentation model is obtained by training a segmentation network using a first training data set; the first training data set includes a plurality of historical egg images with labeled egg pixels and laser line pixels; the egg evaluation model is obtained by training a support vector machine using a second training data set; the second training data set includes egg pixels, laser line pixels of historical egg images and corresponding egg number and egg quality information;

[0034] A feed intake detection module for determining the chicken feed intake according to the feed image and corresponding chicken coop position information; the chicken feed intake is the weight of the chicken feed intake;

[0035] A production performance evaluation module for determining the egg laying hen production performance score by using an egg laying hen production performance evaluation model according to the number of eggs, the egg quality information and the chicken feed intake; wherein the egg laying hen production performance evaluation model is obtained by training a machine learning model using a third training data set; the third training data set includes an input data set and an output data set; the input data set includes the number of eggs, the egg quality information and the chicken feed intake of the training egg laying hens; the output data set is the production performance score of the training egg laying hens given by experts.

[0036] An electronic device, comprising a memory for storing a computer program and a processor for running the computer program to make the electronic device perform the above-mentioned egg laying hen production performance evaluation method.

[0037] According to the specific embodiments of the present application, the following technical effects are disclosed:

[0038] The laying hen production performance evaluation device, method, system and electronic equipment provided by the present application can obtain egg images and feed images through the image acquisition unit of the henhouse inspection module, obtain corresponding position information through the positioning unit of the henhouse inspection module, and determine whether there is an abnormal egg in the cage according to the egg images, so that the henhouse inspection module can be used to perform inspection inside the henhouse, thereby avoiding manual inspection and improving the inspection efficiency. In addition, the host computer can collect egg images and feed images, and process the images to determine the number of eggs, egg quality information and feed intake of the hens. Then, the laying hen production performance is evaluated by using an evaluation model according to the number of eggs, egg quality information and feed intake of the hens, so that the evaluation of the production performance by experts is replaced, the problem of subjective evaluation is avoided, and the accuracy of the evaluation is improved. BRIEF DESCRIPTION OF DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0040] Figure 1 The structure block diagram of the laying hen production performance evaluation device provided by the present application is shown in the figure.

[0041] Figure 2 The schematic diagram of the henhouse inspection module provided by the present application is shown in the figure.

[0042] Figure 3 The schematic diagram of the image acquisition unit provided by the present application is shown in the figure.

[0043] Figure 4 The schematic diagram of the positioning unit provided by the present application is shown in the figure.

[0044] Figure 5 The schematic diagram of another henhouse inspection module provided by the present application is shown in the figure.

[0045] Figure 6 The schematic diagram of another image acquisition unit provided by the present application is shown in the figure.

[0046] Figure 7 The schematic diagram of another positioning unit provided by the present application is shown in the figure.

[0047] Figure 8 The flowchart of the laying hen production performance evaluation method provided by the present application is shown in the figure.

[0048] Figure 9The flowchart for egg quantity and quality detection provided by this invention;

[0049] Figure 10 This invention provides a flowchart for detecting feed intake in chickens.

[0050] Figure 11 A flowchart of the laser three-dimensional measurement method provided by the present invention;

[0051] Figure 12 A schematic diagram of the laser three-dimensional scanning device provided by the present invention;

[0052] Figure 13 This is a schematic diagram of the camera imaging principle provided by the present invention;

[0053] Figure 14 A flowchart for generating the egg production performance distribution map provided by this invention;

[0054] Figure 15 A flowchart illustrating the construction process of the egg production performance evaluation model provided by this invention. Detailed Implementation

[0055] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] The purpose of this invention is to provide a device, method, system, and electronic equipment for evaluating the production performance of laying hens, so as to improve the efficiency and accuracy of evaluating the production performance of laying hens.

[0057] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0058] Example 1

[0059] like Figure 1 As shown, the egg production performance evaluation device provided by the present invention includes: a host computer and a chicken house inspection module (chicken house inspection device).

[0060] The chicken house inspection module comprises n image acquisition units and a positioning unit (positioning device 5); each chicken cage is provided with an image acquisition unit; the image acquisition unit comprises a first camera and a second camera; the first camera is provided with a linear laser for acquiring a feed image; the second camera is provided with a linear laser and a light source for acquiring an egg image; the positioning unit is used to obtain the position of the chicken cage. In actual application, the first camera and the second camera are both near-infrared cameras, and the light of the linear laser and the light source is also near-infrared light. The near-infrared camera adopts a band-pass filter; the imaging wavelength of the near-infrared camera is consistent with the wavelength of the linear laser and the light source.

[0061] The host computer is connected with the image acquisition unit and the positioning unit respectively; the host computer is used to determine the number of eggs and egg quality information according to the egg image, to determine the feed intake of chickens according to the feed image, and to determine the production performance of laying hens according to the number of eggs, the egg quality information and the feed intake of chickens; the egg quality information comprises eggshell quality information, egg geometric shape parameters and egg weight; the eggshell quality information comprises normal eggs, soft-shelled eggs and thin-shelled eggs.

[0062] In actual application, the chicken house inspection module as shown in Figure 2 further comprises a mobile device 1 (mobile unit), a mobile device support 2 (first support), an information acquisition mast 8 (second support), a first fixed block 3 and a second fixed block 7.

[0063] One end of the first support is fixed on the mobile unit; the mobile unit can be a mobile trolley, a feeding cart, a cage door guide rail slider or other mobile platforms; the other end of the first support is provided with the positioning unit; the positioning unit can be an encoder wheel, a visual positioning module or other devices with accurate positioning function; one end of the second support is connected with the first support through the first fixed block 3; the second support is provided with the image acquisition unit.

[0064] One end of the mobile device support 2 is connected with the mobile device 1 at a certain angle, and the other end of the mobile device support 2 is connected with the positioning device 5; the mobile device 1 is used to drive the mobile device support 2 to move in the horizontal direction; the first image acquisition device 4 and the second image acquisition device 6 are connected with the information acquisition mast 8; the first image acquisition device 4 is used to acquire an egg image, and the second image acquisition device 6 is used to acquire a feed image; the information acquisition mast 8 is connected with the mobile device support 2 through the first fixed block 3 and the second fixed block 7; one end of the positioning device 5 is connected with one end of the mobile device support 2, and the other end of the positioning device 5 is connected with the chicken cage crossbar, which is used to determine the position of the chicken cage.

[0065] The target object in the present application is mainly aimed at feed and eggs. The stacked cage livestock house is usually composed of multiple layers of chicken cages arranged in sequence along the vertical direction, and a transverse steel pipe is arranged in the middle part of each side of the stacked chicken cage, which provides conditions for the arrangement and operation of the trolley type inspection device. The trolley type inspection device of the present application is described below.

[0066] Specifically, the mobile device 1 can move autonomously in the chicken house aisle, for example, the mobile device 1 adopts a two-wheeled mobile device, which is simple in structure, small in space occupation, and can adapt to narrow aisles. The mobile device 1 includes a driving wheel and a driven wheel, wherein the driving wheel is a hub motor.

[0067] The mobile device support 2 is installed on the mobile device 1, and the material of the mobile device support can be a stainless steel square tube. When installed, it is at an angle to the vertical direction in order to balance the weight of the information collection mast 8 and prevent the mobile device 1 from tipping over or sliding off the crossbar during travel.

[0068] The information collection mast 8 is installed on the mobile device support 2, and the material of the information collection mast 8 is a carbon fiber square tube or other high-strength square tube. The density is small, which can meet the strength requirements, and at the same time, the weight is light, which is beneficial to reduce the carrying weight of the mobile device 1 and prevent it from rolling over during travel.

[0069] Further, the image acquisition unit comprises a first single-hole expansion seat, a second single-hole expansion seat, a first camera angle adjustment subunit and a second camera angle adjustment subunit.

[0070] The first camera angle adjustment subunit is connected with the first support through the first single-hole expansion seat; the second camera angle adjustment subunit is connected with the second support through the second single-hole expansion seat; the first camera is connected with the first camera angle adjustment subunit; and the second camera is connected with the second camera angle adjustment subunit.

[0071] In actual application, as shown in Figure 3 The first image acquisition device 4 or the second image acquisition device 6 comprises a single-hole expansion seat 41, a camera angle adjustment device 42 and a camera 43. The first image acquisition device 4 is installed on the information collection mast 8, the single-hole expansion seat 41 is connected with the information collection mast 8 and the camera angle adjustment device 42 respectively, which facilitates the movement of the positions of the first image acquisition device 4 and the second image acquisition device 6; the camera 43 (second camera or first camera) is connected with the camera angle adjustment device 42, and the camera can rotate in space, which facilitates the change of the camera shooting range and accurately shoots the eggs or the trough.

[0072] As shown in Figure 4As shown, the positioning device 5 comprises a check ring 51, a fixed sheet 52, a first encoder 53, a fixed rod 54, a torsional spring 55, a first roller 56 and a bearing 57. The positioning device 5 is installed on the mobile device support 2, one end of the fixed rod 54 is connected with one end of the mobile device support 2, the other end of the fixed rod 54 is connected with the fixed sheet 52 through the check ring 51, the torsional spring 55 and the bearing 57, the fixed sheet 52 is connected with the encoder 53, the first encoder 53 is connected with the first roller 56, the first roller 56 is in contact with the chicken cage crossbar at any time, the first roller 56 is made of silica gel and has a circular arc shape, which ensures the stability of the moving direction of the mobile device 1, greatly reduces the risk of roller slipping and the mobile device falling off the crossbar caused by the uneven road surface, and can adapt to various uneven road surfaces.

[0073] The first fixed block 3 and the second fixed block 7 respectively fix two fixed fulcrums of the mobile device support 2, so that the mobile device support 2 is firmly connected with the information collection mast 8, and the stability of image collection is ensured.

[0074] The use process of the present application is described in detail as follows: when the inspection starts, the driving wheel of the mobile device 1 starts to work, driving the mobile device support 2 and the positioning device 5 thereon, and the information collection mast 8 and the first image collection device 4 and the second image collection device 6 thereon to move along the parallel direction of the chicken cage crossbar in the passageway of the livestock and poultry house.

[0075] Specifically, the number of the first image collection device 4 and the second image collection device 6 is the same as the number of layers of the chicken cage, and the camera angle adjusting device 42 is adjusted to adjust the shooting range of the camera 43, so as to complete the monitoring and shooting of the eggs and the trough of each layer of the chicken cage.

[0076] Specifically, in general cases, the fixed sheet 52 is at a certain angle with the vertical direction, and the torsional spring 55 connected with the fixed sheet 52 and the fixed rod 54 generates torsional deformation, so that the roller 56 is always pressed on the chicken cage crossbar to make pure rolling motion, preventing the roller from slipping and ensuring the positioning accuracy, and the rolling of the roller 56 drives the encoder 53 to rotate, thereby recording the position of the target object, i.e. the dead chicken or the egg.

[0077] Specifically, when encountering a bump on the road surface, the fixed rod 54 moves upward with the information collection mast 2, the torsional spring 55 contracts under the action of elastic potential and drives the fixed sheet 52 to rotate along the fixed rod 54 under the action of the bearing 57, the angle between the fixed sheet 52 and the vertical direction is reduced, so that the roller 56 can still be pressed on the chicken cage crossbar; specifically, when encountering a depression on the road surface, the fixed rod 54 moves downward with the information collection mast 2, the torsional spring 55 is twisted and deformed, the elastic potential is enhanced, the force of the roller 56 pressed on the chicken cage crossbar is increased, thereby preventing the roller 56 from slipping and rolling stably on the chicken cage crossbar, and the encoder 53 accurately records the position of the target object, so that the staff can pick up the target object.

[0078] In the embodiment, there is another implementation of the chicken coop inspection module, that is, a trolley adsorption type inspection device.

[0079] As shown in Figure 5 The trolley adsorption type inspection device provided by the embodiment of the application comprises a trolley 10, a chicken coop 9, a third image acquisition device 11, a fourth image acquisition device 12, and a positioning device 13. The trolley 10 is arranged in the chicken coop 9. One end of the third image acquisition device 11 is connected with the trolley 10, and the other end of the third image acquisition device 11 is suspended in the air above an egg belt. The third image acquisition device 11 is used for collecting images of eggs. One end of the fourth image acquisition device 12 is connected with the trolley 10, and the other end of the fourth image acquisition device 12 is suspended in the air above a feed trough. The fourth image acquisition device 12 is used for collecting images of feed. One end of the second positioning device 13 is connected with the trolley 10, and the other end of the second positioning device 13 is connected with the feed trough. The second positioning device 13 is used for determining the position of a target object.

[0080] In the application, the target object mainly refers to the quality and quantity of eggs on the egg belt and the residual amount of feed in the feed trough in the process of raising laying hens. The trolley adsorption type inspection device of the application is described below.

[0081] Specifically, the trolley 10 can move autonomously in a passageway of a livestock and poultry house, and is a device contained in the chicken coop.

[0082] As shown in Figure 6 In the embodiment, the third image acquisition device 11 comprises an image acquisition device fixing block 31, a connecting rod 32, and a camera 34. The third image acquisition device 11 is installed on the trolley 10 through the image acquisition device fixing block 31, and the device is moved along the egg belt transportation direction by the trolley 10. The image acquisition device fixing block 31 can be moved to the positions of the egg belt and the chicken coop at different heights along the vertical direction. The image acquisition device fixing block 31 is made of a magnet, has strong magnetic force, high stability, and is easy to move, so that the position of the third image acquisition device 11 on the trolley can be adjusted conveniently. The image acquisition device fixing block 31 is connected with the connecting rod 32.

[0083] As shown in Figure 7As shown, in this embodiment, the second positioning device 13 includes a positioning device fixing block 61, a bracket 62, a second roller 63, and a second encoder 64. The second positioning device 13 is mounted on the material cart 10 via the positioning device fixing block 61, and is moved along the egg conveyor belt direction by the material cart 10. The positioning device fixing block 61 is made of magnet, which has strong magnetic force, high stability, and is easy to move, facilitating the adjustment of the position of the second positioning device 13 on the material cart 1. The bracket 62 is used to connect and support the positioning device fixing block 61 and the second roller 63. The second roller 63 is made of silicone to reduce friction with the edge of the material trough. The outer contour of the roller is arc-shaped and tangent to the edge of the material trough, ensuring that the roller rolls purely along the edge of the material trough, ensuring stable operation of the positioning device. The second encoder 64 is used to record data, and its recording is accurate, so as to achieve the purpose of collecting the location of the target object.

[0084] The following describes the usage process of this embodiment in detail. The feed cart 10 moves the third image acquisition device 11, the fourth image acquisition device 12 and the second positioning device 13 in the aisle of the livestock and poultry house. The number of chicken house floors to be inspected is changed by changing the position of the fixed block 31 of the image acquisition device. The camera 34 is photographed by changing the camera angle adjustment device 33. The roller 63 drives the encoder 64 to rotate to determine the location of the target object, thereby completing the image acquisition and positioning of abnormal eggs and dead chickens.

[0085] Example 2

[0086] like Figure 8 As shown, this embodiment provides a method for evaluating the production performance of laying hens. This method is applied to the laying hen production performance evaluation device of Embodiment 1, and includes:

[0087] Step 801: Acquire egg images, feed images, and corresponding chicken coop location information; the egg images are acquired using a second camera; the feed images are acquired using a first camera; and the chicken coop location information is acquired using a positioning unit. The egg images include the egg texture outline information under infrared light and the laser line information of the infrared laser.

[0088] Specifically, the second camera and infrared laser emitter (line laser) move above the egg belt in the laying hen house with the moving unit. During the movement, the camera captures images of the eggs at a preset frame rate and uses the positioning unit to obtain the position information of the device when capturing each frame.

[0089] Specifically, the first camera and infrared laser emitter move above the feed trough in the laying hen house with the moving unit. During the movement, feed images are collected at a preset frame rate, and the positioning unit is used to obtain the position information of the device when each frame is collected.

[0090] Step 802: determining the number of eggs and the egg quality information according to the egg image by using an egg detection model; the egg quality information includes eggshell quality information, egg geometric shape parameters and egg weight; the eggshell quality information includes normal eggs, soft-shelled eggs and thin-shelled eggs; wherein the egg detection model includes an egg segmentation model and an egg evaluation model; the egg segmentation model is obtained by training a segmentation network using a first training data set; the first training data set includes a plurality of historical egg images with labeled egg pixels and laser line pixels; the egg evaluation model is obtained by training a model using a second training data set; the second training data set includes egg pixels, laser line pixels of historical egg images and corresponding egg number and egg quality information.

[0091] In practical applications, as shown in Figure 9 The egg evaluation model is used to count the number of eggs and the quality information of each egg according to the egg pixels and laser line pixels in the egg image, combined with the corresponding position information.

[0092] First, the egg image and the corresponding position information under the illumination of the light source and the line laser are acquired in real time by the camera, and a near-infrared light source, a near-infrared laser and a near-infrared camera are used to eliminate the influence of the chicken.

[0093] Specifically, the egg image is acquired in real time, and the corresponding position information is collected at the same time. The egg image and the corresponding position information are collected in real time, so that the eggs in the scanned area on the egg belt can be calculated in real time according to the collected egg image and the corresponding position information.

[0094] Then, based on the egg segmentation model, the egg image is segmented to obtain the egg pixels and the laser line pixels.

[0095] For example, the mask-rcnn instance segmentation network can be used to segment the egg image, which is simple to train and can segment the egg pixels and the laser line pixels in the egg image using a small amount of samples.

[0096] Next, based on the laser three-dimensional measurement method, the three-dimensional point cloud of the egg is calculated according to the laser line pixels; and the egg contour and texture information are obtained according to the egg pixels.

[0097] Finally, based on the egg evaluation model, the number of eggs and the quality information of each egg are determined by comprehensively considering the egg contour, the texture and the three-dimensional point cloud, wherein the egg quality information includes eggshell quality information, egg geometric shape parameters and egg weight.

[0098] For example, the egg evaluation model can use a support vector machine machine learning algorithm, which can solve the classification and regression problems of high-dimensional features, effectively handle nonlinear problems, and accurately predict the egg quality information using the algorithm.

[0099] Step 803: determining the chicken feed intake according to the feed image and the corresponding cage position information; the chicken feed intake is the weight of the feed eaten by the chicken.

[0100] In practical applications, as shown in Figure 10 The laser three-dimensional measurement method is used to obtain the three-dimensional point cloud of the upper surface of the remaining feed in the trough. According to the three-dimensional point cloud of the upper surface of the remaining feed, the shape of the trough and the initial added feed depth are combined to obtain the feed volume eaten by the chicken. According to the density of the feed, the weight of the feed actually eaten by the chicken, i.e. the chicken feed intake, is obtained.

[0101] Further, according to the feed image and the corresponding cage position information, the chicken feed intake is determined, specifically including:

[0102] According to the surface three-dimensional point cloud data, the chicken feed volume is determined in combination with the size of the trough and the initial volume of the feed in the trough. In practical applications, according to the three-dimensional point cloud, the shape of the trough and the initial added feed volume are combined to obtain the feed volume eaten by the chicken. According to the density of the feed, the weight of the feed actually eaten by the chicken, i.e. the chicken feed intake, is obtained.

[0103] According to the surface three-dimensional point cloud data, the chicken feed volume is determined in combination with the size of the trough and the initial volume of the feed in the trough. In practical applications, according to the three-dimensional point cloud, the shape of the trough and the initial added feed volume are combined to obtain the feed volume eaten by the chicken. According to the density of the feed, the weight of the feed actually eaten by the chicken, i.e. the chicken feed intake, is obtained.

[0104] According to the chicken feed volume and the feed density, the chicken feed intake is determined.

[0105] Further, according to the feed image and the corresponding cage position information, the surface three-dimensional point cloud data of the remaining feed in the trough is determined, specifically including:

[0106] The feed image is preprocessed to determine the feed contour and the laser center line; the preprocessing is image segmentation.

[0107] In practical applications, as shown in Figure 11 The infrared laser transmitter emits an infrared laser line, which is irradiated on the measured object.

[0108] Specifically, hens will have a stress reaction to visible light sources, and using ordinary visible light sources will affect the breeding of hens. The wavelength of the near-infrared light or near-infrared laser does not cause stress to the chicken, so the near-infrared light source invisible to the chicken is adopted, which does not disturb the chicken and does not affect production.

[0109] The near-infrared camera is tilted at a certain angle to collect images of the measured object and the near-infrared laser line in real time.

[0110] Specifically, as shown in Figure 12 The near-infrared camera and the infrared laser emitter are connected through a connecting support, and the relative positions are fixed to form a laser three-dimensional scanning device. The laser three-dimensional scanning device is fixed on the moving unit and moves with the moving device, so that the images of the measured object and the infrared laser line can be continuously and real-time collected.

[0111] When collecting images, the infrared fill light is used for lighting, which can clearly collect the contour of the measured object while collecting the laser image. The detection throughput can be improved by increasing the two-dimensional contour of the measured object.

[0112] According to the environmental conditions, the near-infrared camera can be installed with a filter to eliminate the influence of environmental light and increase the stability of three-dimensional measurement.

[0113] The position sensor is used to obtain the position information of the device when collecting images in real time.

[0114] The image is preprocessed, and the preprocessing includes image segmentation, extraction of the contour of the measured object, and extraction of the laser center line.

[0115] For example, if the collected image is an egg image, there will be occlusion between the eggs in actual egg farming. The Mask R-CNN network can be used for instance segmentation of the image, and the egg contour and the laser line contour can be extracted from the instance segmented image. The laser center line can be extracted according to the laser line contour.

[0116] According to the pixel coordinates of the laser center line and the corresponding chicken coop position information, the camera imaging principle is used to determine the surface three-dimensional point cloud data of the remaining feed in the trough.

[0117] Specifically, first, the pixel coordinates of the obtained laser center line are converted into image plane coordinates, and the camera imaging principle is shown in Figure 13 According to the position information of the device and the image plane coordinates of the laser center line, the coordinates of the laser center line in the world coordinate system can be obtained, and the Y coordinate is the position coordinate of the laser three-dimensional scanning device. After processing a series of continuous images obtained by scanning the laser three-dimensional scanning device, the three-dimensional point cloud of the surface of the measured object can be obtained.

[0118] According to the three-dimensional point cloud of the measured object and the contour information of the measured object, the volume of the measured object is obtained by multiple regression.

[0119] Specifically, the two-dimensional contour of the measured object obtained by image segmentation improves the detection flux and improves the accuracy of detection.

[0120] Step 804: According to the egg quantity, the egg quality information and the feed intake of the chicken, an egg production performance evaluation model is used to determine the egg production performance score; wherein the egg production performance evaluation model is obtained by training a machine learning model using a third training data set; the third training data set includes an input data set and an output data set; the input data set includes the egg quantity, the egg quality information and the feed intake of the training egg chicken; and the output data set is the production performance score of the training egg chicken by experts.

[0121] In practical applications, the egg production performance evaluation in the embodiment further includes:

[0122] According to the egg quantity, the egg quality information, the feed intake of the chicken, the cage position information and the egg production performance score, an egg position distribution map, an egg quality information distribution map, a chicken feed intake distribution map and an egg production performance distribution map are generated.

[0123] In one embodiment, as shown in Figure 14 According to the egg quantity, the egg quality information and the feed intake of the chicken, an egg position distribution map, an egg quality information distribution map and a chicken feed intake distribution map are generated, and further based on the egg production performance evaluation model, an egg production performance distribution map is generated, including:

[0124] According to the expert score of the egg production performance, combined with the egg quantity, the egg quality information and the feed intake of the chicken, a machine learning algorithm is used to construct an egg production performance evaluation model.

[0125] Specifically, referring to the egg production performance evaluation model construction flowchart as shown in Figure 15 First, the original data is collected to construct a data set, a batch of egg chickens with different levels of production performance are selected, the expert score of the production performance of the egg chickens is obtained, the number and quality information of the eggs produced by the egg chickens are obtained using the egg quantity and quality detection method, and the feed intake of the chickens is obtained using the chicken feed intake detection method; then the data set is preprocessed using feature engineering method; according to the preprocessed data set, a machine learning algorithm is used to construct an egg production performance evaluation model.

[0126] Based on the egg production performance evaluation model, combined with the real-time collected egg quantity, egg quality information and chicken feed intake, the production performance score of the egg chicken is obtained in real time.

[0127] Specifically, the laying hen production performance evaluation model is deployed on the laying hen production performance evaluation device, the laying hen house to be detected is inspected using the laying hen production performance evaluation device, the number of eggs, egg quality information and the feed intake of the chickens are obtained in real time according to the number of eggs and the quality detection method and the feed intake detection method of the chickens, and the number of eggs, the egg quality information and the feed intake of the chickens are input into the laying hen production performance evaluation model as original data, so that the score of the laying hen production performance can be obtained in real time.

[0128] Based on the position sensor, the corresponding position information is obtained.

[0129] For example, the position sensor can adopt an encoder. The encoder is a device for encoding, converting signals or data into a signal form available for communication, transmission and storage. Using the encoder, the position information of the device can be obtained in real time and accurately, so that the number of eggs, the quality information of the eggs and the feed intake of the chickens can be detected in real time and accurately, and the egg position distribution map, the egg quality information distribution map, the feed intake distribution map and the laying hen production performance distribution map can be generated in real time.

[0130] According to the number of eggs, the egg quality information, the feed intake of the chickens and the corresponding position information, the egg position distribution map, the egg quality information distribution map, the feed intake distribution map and the laying hen production performance distribution map are generated.

[0131] The application provides a laying hen production performance evaluation device, method, system and electronic equipment, and the method comprises the following steps: acquiring an egg image under infrared light irradiation collected by an infrared camera, and acquiring the number of eggs and the quality information of the eggs according to the egg image; acquiring a feed image under infrared laser irradiation collected by the infrared camera, and calculating the residual volume of the feed according to the feed image, so as to obtain the feed intake of the chickens; generating an egg position distribution map, an egg quality information distribution map and a feed intake distribution map according to the number of eggs, the quality information of the eggs and the feed intake of the chickens, and generating a laying hen production performance distribution map based on a laying hen production performance evaluation model. The method has high accuracy in evaluating the production performance of the laying hens, can locate the position of the abnormal chickens in time, can obtain a more accurate laying hen production performance distribution map, and can intuitively and clearly understand the distribution of the production of the laying hens in the hen house.

[0132] Embodiment three

[0133] In order to perform the method corresponding to the above-mentioned embodiment one, to realize the corresponding functions and technical effects, the following provides a laying hen production performance evaluation system, comprising:

[0134] An image acquisition module is configured to acquire an egg image, a feed image, and corresponding cage position information; the egg image is acquired by a second camera; the feed image is acquired by a first camera; and the cage position information is acquired by a positioning unit.

[0135] An egg detection module is configured to determine an egg quantity and egg quality information according to the egg image by using an egg detection model; the egg quality information includes normal eggs, soft-shelled eggs, and thin-shelled eggs; the egg detection model includes an egg segmentation model and an egg evaluation model; the egg segmentation model is obtained by training a segmentation network by using a first training data set; the first training data set includes a plurality of historical egg images with labeled egg pixels and laser line pixels; and the egg evaluation model is obtained by training a support vector machine by using a second training data set; the second training data set includes egg pixels, laser line pixels of historical egg images, and corresponding egg quantity and egg quality information.

[0136] A feed intake detection module is configured to determine a bird feed intake according to the feed image and corresponding cage position information; the bird feed intake is the weight of the feed eaten by the bird.

[0137] A production performance evaluation module is configured to determine an egg-laying hen production performance score by using an egg-laying hen production performance evaluation model according to the egg quantity, the egg quality information, and the bird feed intake; the egg-laying hen production performance evaluation model is obtained by training a machine learning model by using a third training data set; the third training data set includes an input data set and an output data set; the input data set includes the egg quantity, the egg quality information, and the bird feed intake of a training egg-laying hen; and the output data set is a production performance score of the training egg-laying hen given by an expert.

[0138] Embodiment Four

[0139] The present application provides an electronic device, comprising a memory and a processor, the memory is used for storing a computer program, and the processor runs the computer program to make the electronic device execute the egg-laying hen production performance evaluation method of embodiment two.

[0140] In the specification, each embodiment is described in a progressive manner, and each embodiment focuses on the difference from other embodiments, and the same or similar parts between each embodiment can be referred to each other. For the system disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the related parts can be referred to the method part.

[0141] The principles and implementation manners of the present application are described by using specific examples in the present application, and the above examples are only used to help understand the method of the present application and its core idea; meanwhile, for the general technical personnel in the art, the specific implementation manners and application ranges will be changed according to the idea of the present application. In conclusion, the content of the present specification should not be understood as the limitation of the present application.

Claims

1. A device for evaluating the production performance of laying hens, characterized in that, The utility model relates to a kind of automatic egg production performance detection system, including: Host computer and chicken coop inspection module; The chicken coop inspection module includes n image acquisition units and a positioning unit;Each layer of chicken coop is provided with the image acquisition unit;The image acquisition unit includes first camera and second camera;The first camera is used to collect feed image;The first camera is equipped with linear laser;The second camera is used to collect egg image;The second camera is equipped with linear laser and light source;The positioning unit is used to obtain chicken coop position;The light of linear laser and light source is near-infrared light; The host computer is connected with the image acquisition unit and the positioning unit respectively;The host computer is used to determine egg quantity and egg quality information according to the egg image, determine bird intake according to the feed image, determine egg production performance according to the egg quantity, the egg quality information and the bird intake;The egg quality information includes eggshell quality information, egg geometric shape parameter and egg weight;The eggshell quality information normal egg, soft-shelled egg and thin-shelled egg; According to the egg image, the egg quantity and the egg quality information are determined, and specifically include: According to the egg image, egg segmentation model is used to segment egg pixels and laser line pixels; According to the laser line pixels and camera imaging principle, egg three-dimensional point cloud is determined; According to the egg pixel, egg contour and egg texture information are determined; According to the egg contour, the egg texture information and the egg three-dimensional point cloud, egg evaluation model is used to determine egg quantity and egg quality information; According to the feed image, bird intake is determined, and specifically includes: According to the feed image and corresponding chicken coop position information, the surface three-dimensional point cloud data of remaining feed in trough is determined; According to the surface three-dimensional point cloud data, combined with trough size and initial volume of trough feed, bird intake feed volume is determined; According to the bird intake feed volume and feed density, bird intake is determined.

2. The laying hen performance evaluation device according to claim 1, characterized by The first camera and the second camera are near-infrared cameras;The near-infrared camera uses band-pass filter;The imaging wavelength of the near-infrared camera is consistent with the wavelength of linear laser and light source.

3. The laying hen production performance evaluation device according to claim 1, characterized by The chicken coop inspection module further includes a moving unit, a first bracket, a second bracket and a fixing block; One end of the first bracket is fixed to the moving unit;The other end of the first bracket is provided with the positioning unit;One end of the second bracket is connected with the first bracket through the fixing block;The other end of the second bracket is provided with the image acquisition unit.

4. The laying hen production performance evaluation device according to claim 3, characterized by The image acquisition unit includes a first single-hole expansion seat, a second single-hole expansion seat, a first camera angle adjustment subunit and a second camera angle adjustment subunit; The first camera angle adjustment subunit is connected with the first bracket through the first single-hole expansion seat;The second camera angle adjustment subunit is connected with the second bracket through the second single-hole expansion seat;The first camera is connected with the first camera angle adjustment subunit;The second camera is connected with the second camera angle adjustment subunit.

5. A method for evaluating the performance of a laying hen, characterized by, The egg laying performance evaluation method is applied to the egg laying performance evaluation device of any one of claims 1-4, and the egg laying performance evaluation method comprises: acquiring an egg image, a feed image, and corresponding cage position information; the egg image is collected by a second camera; the feed image is collected by a first camera; and the cage position information is collected by a positioning unit; determining the number of eggs and egg quality information according to the egg image by using an egg detection model; the egg quality information includes eggshell quality information, egg geometric shape parameters, and egg weight; the eggshell quality information includes normal eggs, soft-shelled eggs, and thin-shelled eggs; wherein the egg detection model includes an egg segmentation model and an egg evaluation model; the egg segmentation model is obtained by training a segmentation network using a first training data set; the first training data set includes a plurality of historical egg images with labeled egg pixels and laser line pixels; and the egg evaluation model is obtained by training a support vector machine using a second training data set; the second training data set includes egg pixels, laser line pixels, and corresponding egg number and egg quality information of historical egg images; determining the number of eggs and egg quality information according to the egg image by using an egg detection model, specifically comprising: segmenting the egg pixels and laser line pixels according to the egg image by using the egg segmentation model; determining the egg three-dimensional point cloud according to the laser line pixels and the camera imaging principle; determining the egg contour and egg texture information according to the egg pixels; determining the number of eggs and egg quality information according to the egg contour, the egg texture information, and the egg three-dimensional point cloud by using the egg evaluation model; determining the amount of feed consumed by the chicken according to the feed image and the corresponding cage position information; the amount of feed consumed by the chicken is the weight of the feed consumed by the chicken; determining the amount of feed consumed by the chicken according to the feed image and the corresponding cage position information, specifically comprising: determining the surface three-dimensional point cloud data of the remaining feed in the trough according to the feed image and the corresponding cage position information; determining the volume of feed consumed by the chicken according to the surface three-dimensional point cloud data, in combination with the trough size and the initial volume of feed in the trough; determining the amount of feed consumed by the chicken according to the volume of feed consumed by the chicken and the feed density; determining the egg laying performance score of the laying hen by using the egg laying performance evaluation model according to the number of eggs, the egg quality information, and the amount of feed consumed by the chicken; wherein the egg laying performance evaluation model is obtained by training a machine learning model using a third training data set; the third training data set includes an input data set and an output data set; the input data set includes the number of eggs, the egg quality information, and the amount of feed consumed by the chicken of the training laying hen; and the output data set is the production performance score of the training laying hen given by an expert.

6. The method of evaluating production performance of laying hens according to claim 5, wherein, Further comprising: generating an egg position distribution map, an egg quality information distribution map, a chicken feed consumption amount distribution map, and a laying hen production performance distribution map according to the number of eggs, the egg quality information, the amount of feed consumed by the chicken, the cage position information, and the laying hen production performance score.

7. A system for evaluating production performance of a laying hen, characterized by, including: An image acquisition module is configured to acquire an egg image, a feed image, and corresponding cage position information; The egg image is acquired by a second camera; The feed image is acquired by a first camera; and the cage position information is acquired by a positioning unit; An egg detection module is configured to determine an egg quantity and egg quality information according to the egg image by using an egg detection model; the egg quality information includes normal eggs, soft-shelled eggs, and thin-shelled eggs; the egg detection model includes an egg segmentation model and an egg evaluation model; the egg segmentation model is obtained by training a segmentation network by using a first training data set; the first training data set includes a plurality of historical egg images with labeled egg pixels and laser line pixels; and the egg evaluation model is obtained by training a support vector machine by using a second training data set; the second training data set includes egg pixels, laser line pixels of historical egg images, and corresponding egg quantity and egg quality information. The method specifically includes: segmenting egg pixels and laser line pixels from the egg image by using the egg segmentation model; determining an egg three-dimensional point cloud according to the laser line pixels and a camera imaging principle; determining egg contour and egg texture information according to the egg pixels; determining the egg quantity and egg quality information by using the egg evaluation model according to the egg contour, the egg texture information, and the egg three-dimensional point cloud; A feed intake detection module is configured to determine a chicken feed intake according to the feed image and corresponding cage position information; the chicken feed intake is the weight of feed eaten by the chicken. The method specifically includes: determining a surface three-dimensional point cloud data of the remaining feed in the trough according to the feed image and corresponding cage position information; determining a chicken feed intake volume according to the surface three-dimensional point cloud data, in combination with a trough size and an initial volume of feed in the trough; determining the chicken feed intake according to the chicken feed intake volume and a feed density; A production performance evaluation module is configured to determine a layer production performance score by using a layer production performance evaluation model according to the egg quantity, the egg quality information, and the chicken feed intake; the layer production performance evaluation model is obtained by training a machine learning model by using a third training data set; the third training data set includes an input data set and an output data set; the input data set includes an egg quantity, egg quality information, and a chicken feed intake of a training layer; and the output data set is a production performance score of the training layer given by an expert.

8. An electronic device, comprising: The method specifically includes: a memory and a processor; the memory is configured to store a computer program; and the processor is configured to run the computer program to enable the electronic device to perform the layer production performance evaluation method of any one of claims 5-6.

Images