Seed Selection Method and System Based on Machine Vision

Abstract

The present invention relates to the field of seed selection, and particularly to a seed selection method and system based on machine vision. The method includes: obtaining a grayscale image of seeds, extracting the seed region in the grayscale image, where the seed region includes a plurality of first sub-regions, and selecting any pixel point in the first sub-region as a seed point; performing edge detection on the grayscale image to obtain a plurality of edge lines, calculating the internal feature value and the initial edge feature value of each pixel point on the edge line, where the initial feature value is negatively correlated with the internal feature value, and updating the initial edge feature value to obtain the optimal edge feature value; segmenting the grayscale image using the watershed algorithm according to the optimal edge feature value of each pixel point on the edge line to obtain a segmented image; and inputting the segmented image into a detection model to obtain the quality grade of the seeds. The present invention can avoid over-segmentation during the segmentation process, improving the working efficiency and accuracy of seed selection.

Description

Technical Field

[0001] The present invention relates to the field of seed selection, and particularly to a seed selection method and system based on machine vision. Background Art

[0002] Seed selection generally refers to the process of screening out high-quality seeds with high quality, consistent size, and no pests and diseases from a batch of seeds through certain methods and equipment. There are various ways of seed screening, such as screening by seed size and seed weight. When screening rice seeds, by taking pictures of the rice seeds, then extracting the images of the rice seed regions, and inputting the images of the rice seed regions into a convolutional neural network model, the quality of the rice seeds can be detected. For example, the plumpness of the rice seeds, the integrity of the rice seeds, and whether there are pests and diseases can be detected. When extracting the rice seed regions, image segmentation technology is needed to separate the rice seed regions from the background regions, so as to ensure the accuracy of the rice seed detection results.

[0003] The Chinese patent application document with the publication number CN111738256A discloses a method for segmenting and processing CT images of composite materials based on an improved watershed algorithm and morphological evaluation. The method includes: preprocessing the original image to enhance its local features, obtaining initial marker points through morphological processing and distance transformation, and then performing pre-segmentation on the whole by using an adaptive h-value selection algorithm and a watershed algorithm; then evaluating the effectiveness of each connected region in the segmented markers through a regional effectiveness index, and adaptively selecting the local h-value and performing watershed algorithm segmentation on the regions with effectiveness below the set standard, and iterating until almost all connected regions meet the requirements of the effectiveness index, that is, obtaining the final segmentation result of the algorithm.

[0004] In the related art, the watershed algorithm can be used to segment the rice seed regions in the image. However, internal edge lines are formed at the junctions of the glumes on the surface of the rice seeds, and the internal edge lines of the rice seeds have similar characteristics to the contour edge lines of the rice seeds. Therefore, when using the watershed algorithm to segment the image, the problem of over-segmentation is likely to occur, affecting the accuracy of the rice seed detection results. Summary of the Invention

[0005] In order to solve the problem of over-segmentation when using the watershed algorithm to segment images, the present invention provides a seed selection method and system based on machine vision.

[0006] In the first aspect, the present invention provides a seed selection method based on machine vision, adopting the following technical solution:

[0007] Obtain the grayscale image of the seeds, extract the seed regions in the grayscale image. The seed regions include a plurality of first sub-regions, and select any pixel point in the first sub-regions as the seed point;

[0008] Edge detection is performed on a grayscale image to obtain multiple edge lines, and the internal feature value and the initial edge feature value of each pixel point on the edge line are calculated. The initial feature value is negatively correlated with the internal feature value, and the initial edge feature value is updated to obtain the optimal edge feature value;

[0009] The grayscale image is segmented using the watershed algorithm according to the optimal edge feature value of each pixel point on the edge line to obtain a segmented image;

[0010] The segmented image is input into a preset detection model to obtain the quality grade of the seeds;

[0011] Among them, the expression of the internal feature value is: ;

[0012] In the formula, represents the internal feature value of the j-th edge pixel point on the edge line, represents the gradient value of the j-th edge pixel point on the edge line, represents the gray difference value between two symmetric pixel points in the gradient direction of the j-th edge pixel point on the edge line, represents the activation function, and n represents a preset correlation coefficient.

[0013] The optimal edge feature value can reflect whether the corresponding edge pixel point is a contour edge line pixel point or an internal edge line pixel point. By using the optimal edge feature value to replace the gray value in the watershed algorithm, the interference generated by the internal edge line can be weakened during image segmentation, and the phenomenon of over-segmentation can be avoided during the segmentation process, so that a complete rice seed image can be obtained, improving the work efficiency and accuracy of seed selection.

[0014] Preferably, the method for selecting any pixel point in the first sub-region as the seed point is:

[0015] The grayscale image is segmented using the Otsu method to obtain a binary image, the binary image is processed using the distance transformation algorithm to obtain a distance transformation image, and the distance transformation image is binarized to obtain a foreground image and a background image; the foreground image includes multiple second sub-regions, and the pixel point corresponding to the center point of the second sub-region in the grayscale image is used as the seed point.

[0016] The accuracy of the calculation result of the optimal edge feature value is further improved by determining the seed point.

[0017] Preferably, the method further includes: in the grayscale image, calculating the mean value of the gray values of the pixel points corresponding to the foreground image; calculating the background difference feature value of each pixel point in the grayscale image;

[0018] The expression of the background difference feature value is:

[0019] ;

[0020] is the background difference eigenvalue of the i-th pixel point in the corresponding area of the grayscale image and the foreground image. is the absolute value of the difference between the grayscale value of the i-th pixel point in the grayscale image and the mean value in the corresponding area of the foreground image. is the distance between the i-th pixel point in the foreground image and the background.

[0021] The possibility that the corresponding pixel point is an internal point of the rice seed can be preliminarily judged through the background difference eigenvalue, so as to facilitate the distinction between whether the corresponding pixel point belongs to the internal pixel point of the rice seed or the background pixel point.

[0022] Preferably, the expression of the initial edge eigenvalue is:

[0023] ;

[0024] In the formula, is the initial edge eigenvalue of the -th edge pixel point, represents the internal eigenvalue of the j-th edge pixel point, is the background difference eigenvalue of the j-th edge pixel point corresponding in the grayscale image and the foreground image.

[0025] The possibility that the corresponding edge pixel point is a contour edge point can be reflected through the initial edge eigenvalue.

[0026] Preferably, the expression of the optimal edge eigenvalue is:

[0027] ;

[0028] In the formula, is the optimal edge eigenvalue of the -th edge pixel point, is the distance between the m-th edge line where the j-th edge pixel point is located and the seed point, is the -th edge pixel point is located at the -th preset similarity of the edge line, is the initial edge eigenvalue of the -th edge pixel point.

[0029] The possibility that the corresponding edge pixel point is a contour edge line pixel point and an internal edge line pixel point can be accurately reflected through the optimal edge eigenvalue, improving the recognition efficiency of the edge pixel point.

[0030] Preferably, the calculation method of the similarity is:

[0031] Draw a curve of the change of the edge feature value and position of the pixel points on the edge line, and calculate the first similarity between the change curve of any edge line and the change curve of one side edge line; calculate the first distance between the corresponding edge line and the seed point, and the second distance between the one side edge line and the seed point, and take the difference between the first distance and the second distance as the first difference; calculate the second similarity between the change curve of the corresponding edge line and the change curve of the other side edge line, calculate the third distance between the other side edge line and the seed point, and take the difference between the first distance and the third distance as the second difference.

[0032] The expression of the similarity is: ;

[0033] In the formula, is the preset similarity of the th edge pixel point on the th edge line, is the first similarity, is the first difference, is the second similarity, is the second difference.

[0034] Preferably, the detection model is a convolutional neural network model.

[0035] In the second aspect, the present invention provides a seed selection system based on machine vision, adopting the following technical solutions:

[0036] A seed selection system based on machine vision, a processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the seed selection method based on machine vision according to the above is implemented.

[0037] The present invention has the following technical effects:

[0038] Through the optimal edge feature value, it can be reflected whether the corresponding edge pixel point is a contour edge line pixel point or an internal edge line pixel point. By using the optimal edge feature value to replace the gray value in the watershed algorithm, the interference generated by the internal edge line can be weakened when segmenting the image, and the phenomenon of over-segmentation can be avoided during the segmentation process, so that a complete rice seed image can be obtained, improving the working efficiency and accuracy of seed selection. Description of the Drawings

[0039] By reading the following detailed description with reference to the drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become easy to understand. In the drawings, several embodiments of the present invention are shown in an exemplary rather than restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.

[0040] Figure 1It is the flowchart of the seed selection method based on machine vision of the present invention.

[0041] Figure 2 It is the grayscale image of the rice seeds of the present invention. Specific implementation manner

[0042] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts belong to the protection scope of the present invention.

[0043] It should be understood that when the claims, the description and the drawings of the present invention use terms such as "first" and "second", they are only used to distinguish different objects, rather than to describe a specific order. The terms "comprising" and "including" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their combinations.

[0044] Combined with Figure 2 As shown, the surface of the rice seed husk has an edge line. In the horizontal projection of the rice seed, the edge line is located inside the rice seed area. The edge line located inside the rice seed area is used as the internal edge line, and the following will be expanded based on this background.

[0045] The embodiments of the present invention disclose a seed selection method based on machine vision. Refer to Figure 1 , including the following steps, specifically as follows:

[0046] S1: Obtain the grayscale image of the seeds, extract the seed area in the grayscale image. The seed area includes multiple first sub-areas, and any pixel point in the first sub-area is selected as the seed point.

[0047] In one embodiment, the rice seeds to be screened are laid flat on a pure white background board, and a rice seed image is captured from directly above parallel to the background board under uniform illumination conditions. The rice seed image is grayscaled to obtain a grayscale image. The Otsu method is used to segment the grayscale image to obtain a binary image, and the distance transform algorithm is used to process the binary image to obtain a distance transform image. The distance transform image is binarized to obtain a foreground image and a background image, where the grayscale value of the foreground image is 255 and the grayscale value of the background image is 0. The foreground image includes a plurality of second sub-regions, and one second sub-region corresponds to one rice seed. Their distribution state is that adjacent second sub-regions are connected to each other or independent. The geometric center points of the second sub-regions are obtained, and the pixel points corresponding to the center points of the second sub-regions in the grayscale image are used as seed points. It should be noted here that in the grayscale image, the pixel points corresponding to the second sub-regions in the foreground image are the first sub-regions, one first sub-region corresponds to one rice seed, and a plurality of first sub-regions form a seed region.

[0048] In one implementation, in the first sub-region, the grayscale mean value of the pixel points is calculated, and the absolute value of the difference between the grayscale of each pixel point and the grayscale mean value is calculated. The pixel point with the smallest absolute value of the difference is used as the seed point.

[0049] S2: Calculate the background difference eigenvalue of each pixel point in the grayscale image.

[0050] In the grayscale image, the mean value of the grayscale values of the pixel points corresponding to the foreground image is calculated. The expression of the background difference eigenvalue is: ;

[0051] In the formula, is the background difference eigenvalue of the i-th pixel point in the corresponding area of the foreground image in the grayscale image, is the absolute value of the difference between the grayscale value of the i-th pixel point in the corresponding area of the foreground image in the grayscale image and the mean value, is the distance between the i-th pixel point in the foreground image and the background. The distance between the i-th pixel point in the foreground image and the background is calculated by the distance transform algorithm, which is a prior art.

[0052] In the grayscale image, the grayscale values on the surface of the rice seeds are relatively uniform, and the difference from the background grayscale value is large. The larger the value of , the greater the grayscale difference between the corresponding pixel point and the background point, and the greater the probability that it is an internal point of the rice seed. On the contrary, The smaller the value of , the smaller the grayscale difference between the corresponding pixel point and the background point, and the greater the probability that it is a background pixel point;

[0053] In summary, represents the difference feature between the corresponding pixel point in the grayscale image and the background. The larger the value of, the greater the difference between the corresponding pixel point and the background, and the greater the possibility that it is an internal point of the rice seed. The smaller the value of, the smaller the difference between the corresponding pixel point and the background, and the smaller the possibility that it is an internal point of the rice seed.

[0054] S3: Perform edge detection on the grayscale image to obtain multiple edge lines, and calculate the internal feature value and the initial edge feature value of each pixel point on the edge line. The initial feature value is negatively correlated with the internal feature value.

[0055] Use the canny operator to perform edge detection on the grayscale image to obtain the edge lines in the grayscale image. Among them, the edge lines include the contour edge line and the internal edge line of the rice seed, and the internal edge line is the boundary line of the glume. Remove the intersection points between the edge lines, and divide all the edge lines into multiple independent edge segments. On the surface of the rice seed, the color difference between the pixel points on both sides of the internal edge line is small, while the color difference between the pixel points on both sides of the contour edge line of the rice seed is large, that is, the color difference between the pixel points on the surface of the rice seed and the background pixel points is large. Therefore, judge whether it is an internal edge line by the gray value of the pixel points in the normal direction of the edge line.

[0056] S31: Calculate the internal feature value of each pixel point on the edge line.

[0057] The expression of the internal feature value is: ;

[0058] In the formula, represents the internal feature value of the j-th edge pixel point on the edge line, represents the gradient value of the j-th edge pixel point on the edge line, represents the gray difference value of two symmetric pixel points in the gradient direction of the j-th edge pixel point on the edge line. Specifically, it is two symmetric pixel points centered on the j-th edge pixel point. It can be understood that the two pixel points are located in the gradient direction of the j-th edge pixel point. represents the activation function, and n represents the preset correlation coefficient, and its value is 2, so that the value of the internal feature value is in the range of (0, 1).

[0059] The larger the value of, the greater the gray change at the j-th edge pixel point; The smaller the, the smaller the color difference between the two sides of the edge line where the j-th edge pixel point is located, and the greater the possibility that its corresponding edge line is the internal edge line of the rice seed. On the contrary The larger the, the greater the color difference between the two sides of the edge line where the j-th edge pixel point is located, and the greater the possibility that its corresponding edge line is the contour edge line of the rice seed.

[0060] In summary, the internal feature value reflects the internal features of the corresponding edge line. The larger the value, the greater the possibility that the corresponding edge line is an internal edge line. Conversely, the smaller the value, the smaller the possibility that the corresponding edge line is an internal edge line and the greater the possibility that it is a contour edge line.

[0061] S32: Calculate the initial edge feature value of each pixel point on the edge line.

[0062] The expression for the initial edge feature value is: ;

[0063] In the formula, is the initial edge feature value of the th edge pixel point, represents the internal feature value of the jth edge pixel point, is the background difference feature value of the jth edge pixel point corresponding to the foreground image in the grayscale image.

[0064] is the background difference feature value of the th edge pixel point, representing the difference feature between the corresponding pixel point and the background. The larger the value of , the greater the difference between the th edge pixel point and the background, and the greater the possibility that it is an internal point of the rice seed. Conversely, the smaller the value of , the smaller the difference between the th edge pixel point and the background, and it is closer to the contour edge of the rice seed. is the internal edge feature value of the jth edge pixel point. The larger the value of , the greater the possibility that the th edge pixel point is an internal edge point of the rice seed. The smaller the value of , the greater the possibility that the th edge pixel point is a contour edge point of the rice seed.

[0065] Therefore, the possibility that the corresponding edge pixel point is a contour edge point is reflected by the initial edge feature value. Specifically, the larger the value, the more likely the corresponding edge pixel point is a contour edge point of the rice seed, and the smaller the value, the more likely the corresponding edge pixel point is an internal edge point of the rice seed.

[0066] S4: Calculate the similarity of the edge line.

[0067] Take the edge line in the binary image as the baseline. In the grayscale image, remove the edge line corresponding to the baseline. The remaining edge lines are the internal edge lines and the edge lines at the overlapping parts of multiple rice seeds, and calculate the similarity of the remaining edge lines.

[0068] The calculation method is as follows: construct a rectangular coordinate system, with the horizontal axis being the ordinal number of pixel points and the vertical axis being the edge feature value of pixel points. Plot the change curve of the edge feature value and position of pixel points on the edge line in the coordinate system, and use the Fréchet distance algorithm to calculate the first similarity between the change curve of any edge line and the change curve of one side edge line; calculate the first distance between the corresponding edge line and the seed point, and the second distance between one side edge line and the seed point, and take the difference between the first distance and the second distance as the first difference; use the Fréchet distance algorithm to calculate the second similarity between the change curve of the corresponding edge line and the change curve of the other side edge line, calculate the third distance between the other side edge line and the seed point, and take the difference between the first distance and the third distance as the second difference. Among them, the first distance, the second distance, and the third distance are all the minimum distances between the pixel points on the corresponding edge line and the seed point.

[0069] Exemplarily, among the pixel points a, b, c, d on the edge, the distance between point d and the seed point is the smallest. Therefore, the distance between point d and the seed point is taken as the first distance, and the calculation methods of the second distance and the third distance are the same as that of the first distance.

[0070] The expression of similarity is: ; in the formula, is the preset similarity of the th edge pixel point on the th edge line, is the first similarity, is the first difference, is the second similarity, is the second difference.

[0071] The greater the distance difference between the mth edge line and its adjacent edge line to the same seed point, the smaller the possibility that these two edge lines belong to the inner edge line of the same rice seed, and the smaller their similarity; conversely, the smaller the distance difference between these two edge lines to the same seed point, the greater the possibility of belonging to the inner edge line of the same seed, and the greater their similarity.

[0072] is the similarity between the mth edge line where the jth edge pixel point is located and the adjacent two edge lines, The smaller it is, the greater the possibility that the mth edge line is the overlapping outer edge line. Conversely, The greater it is, the greater the possibility that the mth edge line is the inner edge line. It should be noted that when calculating the similarity of edge lines, the edge line corresponding to the baseline has been removed. Therefore, for the similarity of edge lines not participating in the calculation, a fixed value is set according to the actual situation. Exemplarily, the similarity value of edge lines not participating in the calculation is 0.8.

[0073] S5: Update the initial edge feature values to obtain the optimal edge feature values.

[0074] The expression for the optimal edge feature values is:

[0075] ;

[0076] In the formula, is the optimal edge feature value of the j-th edge pixel point, is the distance between the m-th edge line where the j-th edge pixel point is located and the seed point, is the th preset similarity of the edge line where the th edge pixel point is located, is the th initial edge feature value of the

[0077] The smaller the value, the closer the edge line where the j-th edge pixel point is located is to the seed point, and the greater the possibility that it is a pixel point on the internal edge line of the seed. The larger the value, the greater the possibility that the j-th edge pixel point is an external edge point. It should be noted that when calculating the distance between the m-th edge line and the seed point, calculate the distance between each pixel point on the edge line and the seed point, and take the minimum value.

[0078] Through it can accurately reflect the possibility that the j-th edge pixel point is a pixel point on the contour edge line and the internal edge line. Specifically, the larger the value of, the greater the possibility that the j-th edge pixel point is a pixel point on the contour edge line, the smaller the value of, the greater the possibility that the j-th edge pixel point is a pixel point on the internal edge line.

[0079] S6: Use the watershed algorithm to segment the grayscale image based on the optimal edge feature values of each pixel point on the edge line to obtain the segmented image.

[0080] Replace the grayscale value of the pixel point in the watershed algorithm with the optimal feature value of the pixel point, and use the watershed algorithm to segment the grayscale image to obtain the segmented image. One segmented image corresponds to an independent rice seed image.

[0081] S7: Input the segmented image into the preset detection model to obtain the quality grade of the seeds.

[0082] The method for obtaining the detection model is as follows: construct a convolutional neural network model, and train the convolutional neural network model to obtain the detection model. The training method is a prior art and will not be elaborated here. Input the segmented image into the detection model to obtain the quality grade of the seeds. The quality grade is set manually according to the actual situation. Exemplarily, the quality grades include first grade, second grade, and third grade.

[0083] An embodiment of the present invention also discloses a seed selection system based on machine vision, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the seed selection method based on machine vision according to the present invention is implemented.

[0084] The above system further includes other components well-known to those skilled in the art, such as a communication bus and a communication interface. Their settings and functions are known in the art and will not be elaborated here.

[0085] In the present invention, the aforementioned memory can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, device, or component. For example, a computer-readable storage medium can be any suitable magnetic storage medium or magneto-optical storage medium, such as resistive random access memory (RRAM), dynamic random access memory (DRAM), static random access memory (SRAM), enhanced dynamic random access memory (EDRAM), high bandwidth memory (HBM), hybrid memory cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by an application program, module, or both. Any such computer storage medium can be part of the device or accessible or connectable to the device.

[0086] Although this specification has shown and described multiple embodiments of the present invention, it is obvious to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will think of many changes, alterations, and alternative ways without departing from the spirit and idea of the present invention. It should be understood that various alternative solutions to the embodiments of the present invention described herein can be adopted in the process of practicing the present invention.

[0087] The above are all preferred embodiments of the present invention, and the protection scope of the present invention is not limited thereby. Therefore, all equivalent changes made according to the structure, shape and principle of the present invention shall be covered within the protection scope of the present invention.

Claims

1. A seed selection method based on machine vision, characterized in that: Includes steps: Obtaining a grayscale image of the seed, extracting a seed region in the grayscale image, wherein the seed region includes a plurality of first subregions, and selecting any pixel point in the first subregion as a seed point; Perform edge detection on the grayscale image to obtain multiple edge lines, calculate the internal eigenvalue and initial edge eigenvalue of each pixel on the edge line, the initial edge eigenvalue is negatively correlated with the internal eigenvalue, and update the initial edge eigenvalue to obtain the optimal edge eigenvalue; The optimal edge eigenvalue expression is: ; For the The best edge feature value of edge pixels, is the distance between the mth edge line where the jth edge pixel is located and the seed point, For the The edge pixel is located at The preset similarity of the edge lines, For the The initial edge feature value of edge pixels; The similarity calculation method is as follows: draw a change curve of edge feature value and position of pixel points on the edge line, calculate a first similarity between the change curve of any edge line and the change curve of the edge line on one side; calculate a first distance between the corresponding edge line and the seed point, and a second distance between the edge line on one side and the seed point, and use the difference between the first distance and the second distance as the first difference; calculate a second similarity between the change curve of the corresponding edge line and the change curve of the edge line on the other side, calculate a third distance between the edge line on the other side and the seed point, and use the difference between the first distance and the third distance as the second difference; The expression of similarity is: ; is the first similarity, is the first difference, is the second similarity, is the second difference; The grayscale image is segmented using the watershed algorithm according to the optimal edge feature value of each pixel on the edge line to obtain a segmented image; Input the segmented image into a preset detection model to obtain the quality grade of the seeds; The expression of the internal eigenvalue is: ; In the formula, Represents the internal eigenvalue of the jth edge pixel on the edge line, Represents the gradient value of the jth edge pixel on the edge line, Represents the grayscale difference between two pixels that are symmetrical to each other in the gradient direction of the j-th edge pixel on the edge line. represents the activation function, and n represents the preset correlation coefficient.

2. The seed selection method based on machine vision according to claim 1, characterized in that: The method of selecting any pixel point in the first sub-region as a seed point is: The grayscale image is segmented by using the Otsu method to obtain a binary image, the binary image is processed by using the distance transform algorithm to obtain a distance transform image, and the distance transform image is binarized to obtain a foreground image and a background image; The foreground image includes a plurality of second sub-regions, and the pixel points corresponding to the center points of the second sub-regions in the grayscale image are used as seed points.

3. The seed selection method based on machine vision according to claim 2, characterized in that: The method further includes: calculating the mean value of the grayscale values ​​of the pixels corresponding to the foreground image in the grayscale image; calculating the background difference characteristic value of each pixel in the grayscale image; The expression of the background difference eigenvalue is: ； is the background difference feature value of the i-th pixel in the grayscale image and the corresponding area in the foreground image, is the absolute value of the difference between the grayscale value of the i-th pixel in the corresponding area in the foreground image and the mean value, is the distance between the i-th pixel in the foreground image and the background.

4. The seed selection method based on machine vision according to claim 3 is characterized in that: The expression of the initial edge eigenvalue is: ； In the formula, For the The initial edge feature value of edge pixels, represents the internal eigenvalue of the jth edge pixel, is the background difference feature value of the jth edge pixel in the grayscale image corresponding to the foreground image.

5. The seed selection method based on machine vision according to claim 1, characterized in that: The detection model is a convolutional neural network model.

6. The seed selection system based on machine vision is characterized by: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the machine vision-based seed selection method according to any one of claims 1-5 is implemented.

Images