An egg activity automatic detection method and system based on a vision system

Abstract

The application discloses a kind of based on visual system's chicken egg activity automatic detection method and system, method includes the following steps: S1, obtains chicken egg image;S2, the chicken egg image is pretreated, obtains the image that has been pretreated;S3, the chicken egg internal characteristic information of the image that has been pretreated is extracted, obtains input image;S4, by line segment identification algorithm search the texture in the input image, finally obtain the binary image layer containing texture position information;S5, the number of texture in the binary image layer is counted, to determine the activity quality of chicken egg with the number of texture as the basis.The method of the present application identifies the blood silk texture in the egg image through image recognition technology, and determines the activity quality of the measured egg by counting the number of blood silk texture, solves the problem of automatic identification and detection of egg activity quality, liberates mechanical repetitive labor production, improves the accuracy and production efficiency of detection.

Description

Technical Field

[0001] This invention relates to the field of egg activity detection, and more specifically, to an automatic egg activity detection method and system based on a vision system. Background Technology

[0002] Eggs are an indispensable food source in people's daily lives, and their viability is directly related to food safety. Eggs are also a commonly used material in vaccine preparation. In traditional vaccine preparation methods, fertilized live eggs are usually used as a production substrate to culture viruses. The viability of eggs plays a crucial role in vaccine preparation. However, egg quality is affected by many factors, such as differences in egg breed, the content of nutrients within the egg, the age of the egg, and storage conditions, which can affect the cell state within the egg and thus its viability. Therefore, in practical applications, the process from egg cultivation to viral vaccine preparation involves many complex steps. Detecting and classifying egg viability in advance is of great significance for improving vaccine preparation efficiency and egg hatching efficiency.

[0003] Currently, most egg viability testing is done manually. Inspectors shine strong light inside the egg to check for blood vessels and other abnormalities. Manual egg candling is subject to subjective bias, resulting in low efficiency and high repetitiveness. Prolonged repetitive mechanical operation can lead to physical and mental fatigue, causing missed or incorrect detections. Therefore, a vision-based automated egg viability testing method can be developed to alleviate this repetitive mechanical labor, improving accuracy and efficiency.

[0004] The Chinese invention patent application number 202111436545.3, entitled "Intelligent Identification and Detection Method for Individual Egg Quality of Hens," employs the following technical solution: acquiring photos of chicken coops and eggs inside, with the coops marked with a coop number; determining the location of the coop number in the photo; separating the area containing the coop number; segmenting and recognizing the characters of the coop number within the aforementioned area; and using edge detection to identify the eggs in the coop, thereby obtaining the color and quality of the eggs under the corresponding coop number. This technical solution determines egg quality by identifying whether the surface of the tested egg has cracks and damage. The technical solution provided by this invention, however, uses image recognition technology to identify blood vessels in egg images and determines the liveness quality of the tested egg by statistically analyzing the number of blood vessels. Summary of the Invention

[0005] To address the problems of low efficiency, low accuracy, and high repetitiveness in manual inspection of existing technologies, this invention provides an automatic egg activity detection method and system based on a vision system. The technical solution adopted by this invention is as follows:

[0006] The first aspect of this invention provides an automatic egg activity detection method based on a vision system, comprising the following steps:

[0007] S1, Obtain the egg image;

[0008] S2, preprocess the egg image to obtain a preprocessed image;

[0009] S3, extract the internal feature information of the egg from the preprocessed image to obtain the input image;

[0010] S4, the texture in the input image is retrieved by the line segment recognition algorithm, and finally a binarized image layer containing texture location information is obtained;

[0011] S5, count the number of textures in the binarized image layer, and judge the activity quality of the egg based on the number of textures.

[0012] Compared with existing technologies, this invention proposes an automatic egg activity detection method based on a vision system. It identifies blood vessels and textures in egg images through image recognition technology and determines the activity quality of the tested egg by counting the number of blood vessels and textures. This solves the problems of low efficiency, low accuracy, and high repetitiveness of manual detection in existing technologies.

[0013] As a preferred embodiment, in step S2, the preprocessing specifically includes the following steps:

[0014] S21, the ROI region of the egg image is extracted through a pre-set pre-selection box, and the egg image is normalized to obtain a first egg image;

[0015] S22, the first egg image is smoothed using a median filtering algorithm to obtain the second egg image;

[0016] S23, use the Kirchhoff circle detection algorithm to filter out background interference pixels in the second egg image.

[0017] As a preferred embodiment, the specific process of step S23 is as follows: The Kirchhoff circle detection algorithm first uses the gradient method to detect the edges of the second egg image, and then considers the local gradient of each non-zero point in the image. The image gradient is calculated according to a 2×2 mask, using the pixels surrounding the i-th pixel. The angle of its pixel horizontal line is calculated as follows: The gradient magnitude is: Then, according to the analytical equation of a circle: (xa) 2 +(yb) 2 =r 2 Determine the circumference; determine the radius based on the support of all candidate center edges with non-zero pixels; find the Kirchhoff circle with the largest connected component and merge it with the image to filter out light interference around the egg caused by lighting.

[0018] As a preferred approach, in step S3, the feature information inside the egg is extracted by grayscale threshold segmentation and area filtering of the region of interest.

[0019] As a preferred embodiment, step S4 further includes the following steps:

[0020] S41, the texture in the input image is retrieved using a line segment recognition algorithm to obtain an image containing suspected texture;

[0021] S42, use a double threshold method to filter suspected textures in the image, and compare the similarity of the filtered suspected textures by sharpening degree and tilt degree, and stitch adjacent textures with high similarity into the same texture.

[0022] S43, if the stitched texture in step S42 is greater than the set threshold, it is output as the recognized texture, and finally a binarized image layer containing texture location information is obtained.

[0023] As a preferred embodiment, in step S41, the line segment recognition algorithm is specifically implemented as follows:

[0024] The input image in step S3 is used as the algorithm input. The pixels of a single egg in the input image are subjected to Gaussian filtering. A region growing operation is performed on the Gaussian-filtered image. The region growing algorithm is used to merge pixels with similar directions to obtain seed pixels. The seed pixels are used as new region pixels. Each time a new region pixel is added, the angle of the region is updated, thereby obtaining a directional line segment region. The pixels of the line segment region are calculated using rectangle approximation to obtain a regular straight line region. The straight line region is used as the suspected texture.

[0025] As a preferred embodiment, in step S42, the specific implementation process of the dual threshold method is as follows:

[0026] Set a width threshold and a contrast threshold. Filter out some blurry straight line areas based on the contrast threshold, and filter out clear straight line areas that meet the width criteria based on the width threshold.

[0027] As a preferred embodiment, the method further includes step S0, which is performed prior to step S1; step S0 includes:

[0028] S01, Number the image acquisition module and calibrate the camera;

[0029] S02, marking the batch of eggs and different eggs in the same batch;

[0030] S03, Establish a data transmission link.

[0031] A second aspect of the present invention also provides an automatic egg activity detection system based on a vision system, comprising an image acquisition unit, an image preprocessing unit, an image feature extraction unit, a texture recognition unit, and a texture number statistical analysis unit connected in sequence.

[0032] The image acquisition unit is used to acquire an image of an egg;

[0033] The image preprocessing unit is used to preprocess the egg image to obtain a preprocessed image;

[0034] The image feature extraction unit is used to extract the internal feature information of the egg in the preprocessed image to obtain the input image;

[0035] The texture recognition unit is used to retrieve the texture in the input image through a line segment recognition algorithm, and finally obtain a binarized image layer containing texture location information.

[0036] The texture number statistical analysis unit is used to count the number of textures in the binarized image layer, and to determine the activity quality of the egg based on the number of textures.

[0037] A second aspect of the present invention also provides a computer device, including a storage medium, a processor, and a computer program stored in the storage medium and executable by the processor, wherein the computer program, when executed by the processor, implements the steps of the aforementioned automatic egg activity detection method based on a vision system.

[0038] The beneficial effects of this invention are:

[0039] This invention acquires egg images in real time through an image acquisition unit, identifies blood vessels in the egg images using image recognition technology, and determines the activity quality of the tested egg by counting the number of blood vessels. This invention has the advantages of low labor costs, high detection efficiency, and high detection accuracy. Attached Figure Description

[0040] Figure 1 A flowchart of an automatic egg activity detection method based on a vision system provided in an embodiment of the present invention;

[0041] Figure 2 This is a flowchart of the Kirchhoff circle detection process provided in an embodiment of the present invention;

[0042] Figure 3 A flowchart of a suspected texture recognition method is provided for embodiments of the present invention;

[0043] Figure 4 A flowchart of a texture calibration method is provided for embodiments of the present invention;

[0044] Figure 5 A schematic diagram of egg activity quality testing results is provided for an embodiment of the present invention;

[0045] Figure 6 This is a unit connection block diagram of an automatic egg activity detection system based on a vision system provided in an embodiment of the present invention. Detailed Implementation

[0046] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent.

[0047] It should be understood that the described embodiments are merely some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of the embodiments of this application.

[0048] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to limit the embodiments of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0049] In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims. In the description of this application, it should be understood that the terms "first," "second," "third," etc., are used only to distinguish similar objects and are not necessarily used to describe a specific order or sequence, nor should they be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0050] Furthermore, in the description of this application, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. The invention will be further described below with reference to the accompanying drawings and embodiments.

[0051] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0052] Example 1

[0053] Please refer to Figures 1 to 5 An automatic egg activity detection method based on a vision system includes the following steps:

[0054] S1, Obtain the egg image;

[0055] In one specific embodiment, a step S0, performed prior to step S1, is also included; step S0 includes:

[0056] S01, Number the image acquisition module and calibrate the camera;

[0057] Specifically, the camera was calibrated by printing a checkerboard calibration board. Nine calibration experiments were conducted, with the camera shooting distance ranging from 0.2m to 0.8m. Approximately 20 photos were taken from different positions and angles in each group. Black and white feature points in the images were detected. The intrinsic and extrinsic parameters of the camera under ideal distortion-free conditions were calculated using these feature points, and the actual radial distortion was determined using least squares. With tangential distortion: Where r 2 =x 2 +y 2 k1, k2, and k3 are the camera radial distortion parameters, and p1 and p2 are the camera tangential distortion parameters. Based on the actual detection environment, the basic camera parameters are set as follows: acquisition frame rate of 20.16 fps, exposure time of 30 ms, gamma correction of 0.40, balance ratio of 1726, and image gain of 10.

[0058] S02, marking the batch of eggs and different eggs in the same batch;

[0059] Specifically, the batches of eggs are labeled as 'a', where a = 1, 2, 3, ..., z, and z is a positive integer. Different eggs within the same batch are labeled as 'egg', where egg = 1, 2, ..., x, and x is a positive integer. Multiple cameras are used to capture the quality of eggs from the same batch. Each camera captures two rows of eggs, ensuring the clarity of the blood vessels and texture while minimizing pixel noise caused by uneven lighting. The texture features of both sides of the same egg are tested as the basis for judgment, ensuring no missed, incorrect, or misjudged cases.

[0060] S03, Establish the data transmission link;

[0061] Specifically, the image acquisition module inputs images and current camera parameter information through multiple threads, and transmits data via a switch connected to the PC. Synchronous transmission of images from multiple cameras requires packet switching technology from the switch, which physically connects to each network node via fiber optic cables. The switch uses routing protocols to determine the data transmission path based on IP addresses and subnet masks, improving the efficiency of multi-camera data transmission.

[0062] Specifically, in step S1, images of the same egg are captured from two different angles. The eggs captured by different cameras are numbered according to a set pattern, and the same number is assigned to images of the same egg captured by different cameras.

[0063] S2, preprocess the egg image to obtain a preprocessed image;

[0064] In one specific embodiment, step S2, the preprocessing specifically includes the following steps:

[0065] S21, the ROI region of the egg image is extracted through a pre-set pre-selection box, and the egg image is normalized to obtain a first egg image;

[0066] S22, the first egg image is smoothed using a median filtering algorithm to obtain the second egg image;

[0067] S23, use the Kirchhoff circle detection algorithm to filter out background interference pixels in the second egg image.

[0068] In a specific embodiment, the specific process of step S23 is as follows: The Kirchhoff circle detection algorithm first uses the gradient method to detect the edges of the second egg image, and then considers the local gradient of each non-zero point in the image. The image gradient is calculated according to the mask 2×2, using the pixels around the i-th pixel. The angle of its pixel horizontal line is calculated as follows: The gradient magnitude is: Then, according to the analytical equation of a circle: (xa) 2 +(yb) 2 =r 2 Determine the circumference; determine the radius based on the support of all candidate center edges with non-zero pixels; find the Kirchhoff circle with the largest connected component and merge it with the image to filter out light interference around the egg caused by lighting.

[0069] S3, extract the internal feature information of the egg from the preprocessed image to obtain the input image;

[0070] In a specific embodiment, step S3 involves extracting the feature information inside the egg by grayscale threshold segmentation and area filtering of the region of interest.

[0071] S4, the texture in the input image is retrieved by the line segment recognition algorithm, and finally a binarized image layer containing texture location information is obtained;

[0072] In one specific embodiment, step S4 further includes the following steps:

[0073] S41, the texture in the input image is retrieved using a line segment recognition algorithm to obtain an image containing suspected texture;

[0074] S42, use a double threshold method to filter suspected textures in the image, and compare the similarity of the filtered suspected textures by sharpening degree and tilt degree, and stitch adjacent textures with high similarity into the same texture.

[0075] S43, if the stitched texture in step S42 is greater than the set threshold, it is output as the recognized texture, and finally a binarized image layer containing texture location information is obtained.

[0076] It should be noted that, in step S41, the specific implementation steps of the line segment recognition algorithm are as follows:

[0077] The input image in step S3 is used as the algorithm input. The pixels of a single egg in the input image are subjected to Gaussian filtering. A region growing operation is performed on the Gaussian-filtered image. The region growing algorithm and the gradient magnitude from step S23 are used to merge pixels with similar directions to obtain a seed pixel. The seed pixel is used as a new region pixel. Each time a new region pixel is added, the angle of the region is updated, thus obtaining a directional line segment region. The pixels of the line segment region are approximated using rectangles to obtain a regular straight line region. The straight line region is used as the suspected texture.

[0078] It should be noted that the straight-line region is an approximately rectangular region composed of several pixels, and this region can be regarded as a straight line under the conventional image size.

[0079] Specifically, in the step S41, when using the rectangular approximation to calculate the pixels in the line segment region, the angle of the rectangle is set to the angle of the eigenvector, and this eigenvector is related to the minimum eigenvalue of the M matrix: Where G i is the gradient value of pixel i, and i represents each pixel point in the region. (c x ,c y ) is the center of the rectangle.

[0080] It should be noted that in the step S42, the specific implementation process of the double-threshold method is as follows:

[0081] Set the width threshold and the contrast threshold. Filter out some blurred straight-line regions according to the contrast threshold, and screen out the clear straight-line regions that meet the width conditions as textures through the width threshold.

[0082] Specifically, the set width threshold and contrast threshold divide the gradient magnitude into two parts. Those greater than the high threshold P1 are considered strong textures, those less than the low threshold P2 are considered non-textures (noises), and those between the two thresholds are considered weak textures P2 < P ≤ P1. Then, in the weak textures, detect and determine whether there is pixel information connected to the strong textures. If so, classify them into the same type of texture information, otherwise filter them out. Finally, mark the suspected textures.

[0083] S5. Count the number of textures in the binary image layer, and judge the activity quality of the eggs based on the number of textures.

[0084] Embodiment 2

[0085] Please refer to Figure 6 , an automatic egg activity detection system based on a vision system, including an image acquisition unit 1, an image preprocessing unit 2, an image feature extraction unit 3, a texture recognition unit 4, and a texture number statistical analysis unit 5 connected in sequence;

[0086] The image acquisition unit 1 is used to acquire an egg image;

[0087] The image preprocessing unit 2 is used to preprocess the egg image to obtain a preprocessed image;

[0088] The image feature extraction unit 3 is used to extract the internal feature information of the eggs in the preprocessed image to obtain an input image;

[0089] The texture recognition unit 4 is used to retrieve the texture in the input image through a line segment recognition algorithm, and finally obtain a binarized image layer containing texture location information.

[0090] The texture number statistical analysis unit 5 is used to count the number of textures in the binarized image layer and to determine the activity quality of the egg based on the number of textures.

[0091] Example 3

[0092] A computer device includes a storage medium, a processor, and a computer program stored in the storage medium and executable by the processor, wherein the computer program, when executed by the processor, implements the steps of the automatic egg activity detection method based on a vision system in Embodiment 1.

[0093] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for automatic detection of egg activity based on a vision system, characterized in that, The method comprises the following steps: S1, acquiring an egg image; S2, preprocessing the egg image to obtain a preprocessed image; S3, extracting internal feature information of the preprocessed image to obtain an input image; S4, searching for textures in the input image by using a line segment recognition algorithm to obtain a binary image layer containing texture position information; S5, counting the number of textures in the binary image layer, and judging the activity quality of the egg according to the number of textures; In the step S4, the following steps are further included: S41, searching for textures in the input image by using a line segment recognition algorithm to obtain an image containing suspected textures; S42, screening the suspected textures in the image by using a double-threshold method, and comparing the similarity of the sharpening degree and the inclination degree of the screened suspected textures, and splicing adjacent textures with higher similarity as the same texture; S43, if the spliced texture in the step S42 is greater than a set threshold, the texture is output as a recognized texture, and finally a binary image layer containing texture position information is obtained; In the step S41, the line segment recognition algorithm specifically comprises the following steps: The input image in the step S3 is taken as an algorithm input, the pixels of a single egg in the input image are subjected to Gaussian filtering, the image subjected to the Gaussian filtering is subjected to region growing operation, and seed pixels are obtained by using a region growing algorithm to merge pixels with similar directions, the seed pixels are taken as new region pixels, and the angle of the region is updated every time a new region pixel is added, so that a directional line segment region is obtained; the pixels in the line segment region are calculated by using rectangular approximation to obtain a regular straight line region; and the straight line region is taken as the suspected texture; The angle of the rectangle is set as the angle of an eigenvector related to the minimum eigenvalue of an M matrix; the M matrix is specifically: ; wherein, , , ; wherein, represents the gradient value of the pixel i, i represents each pixel point in the region, represents the center of the rectangle.

2. The method for automatic detection of egg freshness based on vision system according to claim 1, characterized in that, In the step S2, the preprocessing specifically comprises the following steps: S21, extracting an ROI region of the egg image by using a pre-set preselection frame, and performing normalization processing on the egg image to obtain a first egg image; S22, performing smoothing processing on the first egg image by using a median filter algorithm to obtain a second egg image; S23, filtering background interference pixels of the second egg image by using a Kirchhoff circle detection algorithm.

3. The method of claim 2, wherein the method is a method of automatic detection of egg freshness based on a vision system. The specific process of the step S23 is as follows: the Kirchhoff circle detection algorithm firstly detects the edge of the second egg image by using the gradient method, and then considers the local gradient of each non-0 point of the image, the image gradient is calculated according to the mask 2*2, and the pixels around the i-th pixel point are used for calculation: , ; The pixel horizontal line angle is calculated as: The gradient amplitude is: According to the analytic equation of the circle: The circumferential line is determined; the radius is determined according to the support degree of the non-0 pixels of the edge on all candidate centers; the maximum connected domain of the Kirchhoff circle is found out and merged with the image, so as to filter the light interference around the egg due to the light.

4. The method for automatic detection of egg freshness based on vision system according to claim 1, characterized in that, In the step S3, the internal feature information of the egg is extracted by using gray threshold segmentation and area screening of a region of interest.

5. The method for automatic detection of egg freshness based on vision system according to claim 1, characterized in that, In the step S42, the double-threshold method specifically comprises the following steps: The width threshold and the contrast threshold are set, the contrast threshold is used to filter out part of the fuzzy straight line region, and the width threshold is used to screen clear straight line regions meeting the width condition.

6. The method for automatic detection of egg freshness based on vision system according to claim 1, wherein, The step S0 executed before the step S1 is further included; the step S0 comprises: S01, numbering an image acquisition module and calibrating a camera; S02, marking batches of eggs and different eggs in the same batch; S03, building a data transmission link.

7. The method of claim 6, wherein the method is a method of automatic detection of egg freshness based on a vision system. In the step S01, the specific process of the camera calibration comprises: The printed chessboard calibration board is used to set up several groups of calibration experiments, the camera shooting distance is within a preset range, N pictures of different positions and angles are shot for each group, where N is a positive integer; black and white feature points in the images are detected, camera internal parameters and external parameters under ideal non-distortion conditions are solved through the feature points, and actual radial distortion and tangential distortion are solved through least squares: ; and tangential distortion: ; basic parameters of the camera are set according to the radial distortion and the tangential distortion, wherein, and respectively represent normalized horizontal coordinates and vertical coordinates of the points after distortion, is a radial distortion parameter of the camera, is a tangential distortion parameter of the camera.

8. An automatic egg activity detection system based on a vision system, characterized by, The system comprises sequentially connected image acquisition unit (1), image preprocessing unit (2), image feature extraction unit (3), texture recognition unit (4) and texture number statistical analysis unit (5); The image acquisition unit (1) is used for acquiring egg images; The image preprocessing unit (2) is used for preprocessing the egg images to obtain preprocessed images; The image feature extraction unit (3) is used for extracting internal feature information of the preprocessed images to obtain input images; The texture recognition unit (4) is used for searching textures in the input images by using a line segment recognition algorithm to finally obtain a binary image layer containing texture position information; The texture number statistical analysis unit (5) is used for counting the number of textures in the binary image layer to determine the activity quality of the egg according to the number of textures; The line segment recognition algorithm is used to search the textures in the input images to obtain images containing suspected textures; The suspected textures in the images are screened by using a double-threshold method, and the screened suspected textures are compared in terms of the similarity of sharpening degree and inclination degree, and the adjacent textures with high similarity are spliced into the same texture; If the spliced texture is greater than a set threshold, it is output as a recognized texture, and finally a binary image layer containing texture position information is obtained; The line segment recognition algorithm comprises the following steps: An input image is taken as an algorithm input, and the pixels of a single egg in the input image are subjected to Gaussian filtering, and the image subjected to Gaussian filtering is subjected to region growing operation, and the region growing algorithm is used to combine the pixels with similar directions to obtain seed pixels, the seed pixels are taken as new region pixels, the angle of the region is updated every time a new region pixel is added, and thus a directional line segment region is obtained; the pixels in the line segment region are calculated by using rectangular approximation to obtain a regular straight line region; and the straight line region is taken as the suspected texture. The angle of the rectangle is set as the angle of an eigenvector related to the minimum eigenvalue of an M matrix; the M matrix is specifically: ; wherein, , , ; wherein, indicates the gradient value of pixel i, i indicates each pixel point in the region, indicates the center of the rectangle.

Images