A non-destructive testing method for the angle of eggplant stems

Abstract

This invention discloses a non-destructive detection method for eggplant stem angles, comprising the following steps: S1 Input a captured image of the eggplant stem angle. S2 Perform Hough line detection on the image. S3 Determine the number of successfully combined rectangles. S4 Combine all rectangles into a new contour. S5 Segment the background and eggplant stem images according to the minimum bounding rectangle and convert the images to grayscale. S6 Binarize the image using the large law method. S7 Skeletonize the binarized image to obtain an eggplant stem skeleton map. S8 Analyze the eggplant stem skeleton map and perform corner detection. All judgment thresholds in this invention are automatically generated after analyzing the eggplant stem angle image, and the generated results are unique.

Description

Technical Field

[0001] This invention belongs to the field of crop phenotyping, and more specifically, relates to a non-destructive method for detecting the angle between eggplant stems. Background Technology

[0002] Since the early 1990s, "omics" has become one of the most prominent new concepts and disciplines in the life sciences, including genomics, transcriptomics, proteomics, and metabolomics. With the continuous rise and development of various omics disciplines, the concepts of phenome and phenomics were proposed in the late 1990s. A phenome refers to all the phenotypic characteristics of an organism; phenomics is a discipline that systematically studies all phenotypes of an organism or cell under various environmental conditions at the genomic level. Phenomics, leveraging high-throughput phenotypic analysis technologies and platforms, has combined with genomics, transcriptomics, proteomics, and metabolomics, and has become a major technological platform in systems biology, applied to the study of complex living systems. In the past decade or so, phenomics and phenotyping research platforms, which were originally mainly used in animal and biomedical research, have also been incorporated into plant phenotyping and related omics research. Undoubtedly, phenotyping, as a research field related to phenotyping identification, is a bridge connecting the genotype and phenotype of an organism; at the same time, phenotyping will provide a basis for functional genomics research by utilizing detailed phenotyping information of the target population.

[0003] High-throughput phenomics research typically requires measuring a large number of samples in a very short time, necessitating various devices and algorithms to acquire the required phenotypic features. CN107507192A, "A Portable Device and Method for In Vivo Measurement of Rice Flag Leaf Angle," mentions a method for measuring rice flag leaf angle, but this method is not suitable for the inflexible eggplant stem, and its algorithm is also unsuitable for the rough skin of the eggplant stem. Wang Tao, Gong Liang, et al., in their paper "Measurement of Rice Flag Leaf Angle Based on Custom Clustering," mention a method that effectively measures the angle of rice flag leaves, but this method is not applicable to the eggplant breeding stage, where plants are generally grown in greenhouses, where the white background is difficult to separate from the white greenhouse film. Furthermore, while rice flag leaves have distinct features for location, eggplant stems lack such features. Therefore, proposing a non-destructive detection method for eggplant stem angle to quickly and stably calculate the angle and avoid subjective errors caused by manual measurement is urgently needed. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a non-destructive testing method for the angle of eggplant stems.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a non-destructive testing method for the angle of an eggplant stem, comprising the following steps:

[0006] S1. Input the captured image of the angle between the eggplant stems;

[0007] S2. Perform Hough line detection on the image, attempt to combine the detected lines into rectangles, and record all successfully combined rectangles.

[0008] S3. Determine the number of successfully combined rectangles. If the number is greater than N (N>5 by default), proceed to step S4. If the number is less than 5, the image analysis is considered to have failed.

[0009] S4. Combine all rectangles into a new outline and calculate the minimum bounding rectangle of the outline.

[0010] S5. Divide the background and eggplant stem images into grayscale by using the smallest bounding rectangle.

[0011] S6. Use the big rule method to binarize the image and segment out the eggplant stem.

[0012] S7. Perform a skeletonization operation on the binarized image to obtain the eggplant stem skeleton map.

[0013] S8. Analyze the eggplant stem skeleton diagram, perform corner detection, and obtain key nodes of the skeleton diagram. At the same time, determine whether the key nodes include the bottommost point of the eggplant stem within the calibration board. If not, add the bottommost point to the key nodes.

[0014] S9. Summarize the key nodes and draw the straight line L from the bottom to the fork point. b The straight line L from the bifurcation point to the nearest left or right side point l and L r Calculate L l and L r Angle.

[0015] S10, Output Results

[0016] Preferably, in step S2, the acquired image of the eggplant stem angle should include the complete background board and the plant roots to be detected. During shooting, the imaging device should be centered on and parallel to the background board as much as possible. The background board is printed with a hollow checkerboard pattern.

[0017] Preferably, in step S2, the Hough line detection is implemented using the image processing library Empusv, with the following parameters:

[0018] Preferably, in step S4, the process of combining all rectangles into a new contour only involves filling the new contour with the vertices of each rectangle, thereby reducing the amount of computation in the system.

[0019] Preferably, in step S6, the large law method performs a Gaussian filtering operation on the image before binarization, and the default Gaussian kernel size is H. e = w * 2 + 1, where w is the width of the checkerboard boundary, implemented by the function Threshold() of the image processing library Emcuv.

[0020] Preferably, in step S8, corner detection involves traversing all points on the skeletal line and determining the grayscale similarity between all points in the vicinity of a given point and that point.

[0021] Where (x,y) and (x0,y0) are the coordinates of the center pixel of the template and the pixel to be compared, respectively, and t is the similarity threshold. Furthermore, since the formula does not consider the case where (x,y) is at the image boundary, it is necessary to check whether the corner points contain the bottom of the skeletal line; if not, it is included.

[0022] Preferably, in step S9, the bottom straight line L is formed according to the positional relationship of the points. b Used for positioning, left-side branch L l and the right L R Used to calculate the angle between eggplant stems.

[0023] The technical effects and advantages of this invention are as follows: All judgment thresholds in this invention are automatically generated after analyzing the eggplant stem angle image, and the generated results are unique. This effectively reduces the difficulties caused by inconsistent standards and huge workload due to manual intervention, laying a good foundation for subsequent work. Furthermore, the calculation process of this algorithm is simple, which improves the speed of measuring the eggplant stem angle to a certain extent, thereby providing more stable and efficient data support for eggplant phenotypic research. Attached Figure Description

[0024] Figure 1 This is a schematic diagram illustrating the shooting requirements of the present invention;

[0025] Figure 2 This is a schematic diagram of the image processing flow results of the present invention;

[0026] Figure 3 This is a flowchart of the algorithm of the present invention; Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without inventive effort are within the scope of protection of this invention.

[0028] This invention provides a technical solution: an automatic segmentation algorithm for edible mushroom cluster images, comprising the following steps:

[0029] S1. Input the captured image of the angle between the eggplant stems;

[0030] S2. Perform Hough line detection on the image, attempt to combine the detected lines into rectangles, and record all successfully combined rectangles.

[0031] S3. Determine the number of successfully combined rectangles. If the number is greater than N (N>5 by default), proceed to step S4. If the number is less than 5, the image analysis is considered to have failed.

[0032] S4. Combine all rectangles into a new outline and calculate the minimum bounding rectangle of the outline.

[0033] S5. Divide the background and eggplant stem images into grayscale by using the smallest bounding rectangle.

[0034] S6. Use the big rule method to binarize the image and segment out the eggplant stem.

[0035] S7. Perform a skeletonization operation on the binarized image to obtain the eggplant stem skeleton map.

[0036] S8. Analyze the eggplant stem skeleton diagram, perform corner detection, and obtain key nodes of the skeleton diagram. At the same time, determine whether the key nodes include the bottommost point of the eggplant stem within the calibration board. If not, add the bottommost point to the key nodes.

[0037] S9. Summarize the key nodes and draw the straight line L from the bottom to the fork point. b The straight line L from the bifurcation point to the nearest left or right side point l and L R Calculate L l and L R Angle.

[0038] S10, Output Results

[0039] Specifically, the eggplant stem angle image obtained in the non-destructive testing method S1 for eggplant stem angle should include, for example: Figure 1 As shown: a complete background board and the plant roots to be inspected. During photography, the imaging device should be centered and parallel to the background board as much as possible. The background board is printed with a hollow checkerboard pattern.

[0040] Specifically, a non-destructive testing method for the angle between eggplant stems, such as... Figure 2 b The image is subjected to Hough line detection. The detected lines are combined to attempt to form rectangles, and all successfully combined rectangles are recorded.

[0041] Specifically, a non-destructive testing method for the angle between eggplant stems, such as... Figure 2 b The diagram shows how to combine all rectangles into a new contour, and then fill the new contour with only the vertices of each rectangle, reducing the amount of computation in the system. At the same time, it determines the plane where the eggplant stem is located, improving the calculation accuracy.

[0042] Specifically, a non-destructive testing method for the angle between eggplant stems, such as... Figure 2 C The minimum bounding rectangle operation is performed on the new contour formed by all the rectangles to determine the plane in which the eggplant stem is located, thus improving the calculation accuracy.

[0043] Specifically, a non-destructive testing method for the angle between eggplant stems, such as... Figure 2 d The method shown cuts out the plane containing the eggplant stem from the original image and performs grayscale conversion, reducing the amount of computation required by the system.

[0044] Specifically, a non-destructive testing method for the angle between eggplant stems, such as... Figure 2 e The image shown is binarized using the large-scale law method after grayscale conversion, eliminating the checkerboard pattern and segmenting the binary image of the eggplant stem region.

[0045] Specifically, a non-destructive testing method for the angle between eggplant stems, such as... Figure 2 f The binary image of the segmented eggplant stem region is shown to be skeletalized to obtain the skeletal image of the eggplant stem region.

[0046] Specifically, a non-destructive testing method for the angle between eggplant stems, such as... Figure 2 g Corner point determination is performed on the points on the skeletal image of the eggplant stem region shown, and the key point set of the skeletal image of the eggplant stem region is extracted. Due to the limitations of the corner point determination algorithm, before proceeding to the next step, all points in the skeletal image of the eggplant stem region need to be sorted, the bottommost point needs to be identified, and it needs to be determined whether this point exists in the key point set. If it does not exist, it is added to the key point set.

[0047] Specifically, a non-destructive testing method for the angle between eggplant stems, such as... Figure 2 h The positions of the points shown form the bottom straight line L. b Used for positioning, left-side branch L l and the right L r This is used to calculate the angle between eggplant stems. The final calculation result is then output to the original image for display.

[0048] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A non-destructive testing method for the angle between eggplant stems, characterized in that: Includes the following steps: S1. Input the captured image of the angle between the eggplant stems; S2. Perform Hough line detection on the image, attempt to combine the detected lines into rectangles, and record all successfully combined rectangles. S3. Determine the number of successfully combined rectangles. If the number is greater than N (N>5 by default), proceed to step S4. If the number is less than N, the image analysis is considered to have failed. S4. Combine all rectangles into a new outline and calculate the minimum bounding rectangle of the outline. S5. Divide the background and eggplant stem images into grayscale by using the smallest bounding rectangle; S6. Use the big rule method to binarize the image and segment out the eggplant stem; S7. Perform a skeletonization operation on the binarized image to obtain the eggplant stem skeleton map; S8. Analyze the eggplant stem skeleton diagram, perform corner detection, and obtain key nodes of the skeleton diagram; at the same time, determine whether the key nodes include the bottom point of the eggplant stem in the calibration plate. If not, add the bottom point to the key nodes. S9. Summarize the key nodes and draw the straight line L from the bottom to the fork point. b The straight line L from the bifurcation point to the nearest left or right side point l and L r Calculate L l and L r included angle; S10, Output the result.

2. The non-destructive testing method for the angle of eggplant stems according to claim 1, characterized in that: In step S1, the acquired image should include the complete background board and the plant roots to be detected; during shooting, the imaging device should be centered on the background board and parallel to the background board as much as possible; the background board is printed with a hollow checkerboard pattern.

3. The non-destructive testing method for the angle of an eggplant stem according to claim 1, characterized in that: In S2, Hough line detection is implemented using the image processing library Emgusv.

4. The non-destructive testing method for the angle of eggplant stems according to claim 1, characterized in that: In step S4, during the process of combining all rectangles into a new contour, only the vertices of each rectangle are used to fill the new contour, reducing the amount of computation in the system.

5. The non-destructive testing method for the angle of an eggplant stem according to claim 1, characterized in that: In step S6, the large law method binarizes the image using the function Threshold() from the image processing library Emcuv. Before binarization, image scaling or filtering operations can be added to reduce the influence of the checkerboard pattern in the background.

6. The non-destructive testing method for the angle of an eggplant stem according to claim 1, characterized in that: In step S8, corner detection is obtained by comparing the grayscale similarity between all points in the vicinity of the determination point and that point. Where (x,y) and (x0,y0) are the coordinates of the center pixel of the template and the pixel to be compared, respectively, and t is the similarity threshold; in addition, the bottom point of the inner eggplant stem is obtained by taking the last point after sorting all points in ascending order of Y-axis coordinates.

7. The non-destructive testing method for the angle of an eggplant stem according to claim 1, characterized in that: In S9, where L b L l and L r All are automatically generated by the system and are unique.

Images