Algae phenotypic character measuring method and measuring device

Abstract

The invention belongs to the technical field of algae phenotypic character measurement, and particularly relates to a kelp phenotypic character measurement method and device, and the method comprises the steps: extracting feature points of kelp from each local image of the kelp; for each feature point, calculating a descriptor of a local area where the feature point is located; calculating the similarity of descriptors among the continuous kelp local images, and finding matched feature point pairs among the continuous kelp local images; after the matched feature point pairs are found, calculating a homography matrix between the two continuous local kelp images, and performing perspective transformation to obtain a preliminarily spliced complete kelp image; performing image fusion on the preliminarily spliced complete kelp image; and real kelp phenotype data is obtained by taking the calibration plate as a calculation template through a segmentation algorithm. According to the method, the phenotypic characters of the kelp can be rapidly and accurately measured.

Description

Technical Field

[0001] The invention belongs to the technical field of algae phenotypic trait measurement, and particularly relates to a method and a device for measuring kelp phenotypic traits. Background Art

[0002] Traditionally, the phenotypic traits of large kelp are measured manually, which can achieve a certain degree of accuracy when measuring the length and width of seaweed.

[0003] However, during manual measurement, the accuracy of the measurement results may be affected due to factors such as the operator's skill level, experience differences, and visual fatigue. The accuracy of manual measurement still needs to be improved, and the measurement speed is slow and the efficiency is low. Summary of the Invention

[0004] In order to solve the technical problems that "manual measurement of seaweed phenotypic traits still needs to be improved in accuracy, and the measurement speed is slow and the efficiency is low".

[0005] The present invention provides the following technical solutions:

[0006] Firstly,

[0007] The present invention provides a method for measuring phenotypic traits of kelp, comprising the following steps:

[0008] Step 1: Place the calibration plate and kelp flat on the stage, flatten the kelp, and take multiple partial images of the kelp;

[0009] Step 2: Extract the feature points of kelp from each local image of kelp; for each feature point, calculate the descriptor of the local area where it is located;

[0010] Calculate the similarity of descriptors between consecutive kelp local images; find matching feature point pairs between consecutive kelp local images based on the similarity of descriptors;

[0011] Step 3: After finding the matching feature point pairs, calculate the homography matrix between the two consecutive kelp local images;

[0012] Step 4: After the homography matrix is ​​calculated, the perspective transformation is performed on the local kelp image so that two consecutive local kelp images are aligned in the same coordinate system to achieve the kelp image registration effect and obtain a preliminary spliced ​​complete kelp image;

[0013] Step 5: Perform image fusion on the preliminarily stitched complete kelp images to generate a smooth complete kelp image;

[0014] Step 6: Use the segmentation algorithm to extract the kelp sample area in the smooth complete kelp image and calculate the number of pixels it occupies; use the calibration plate as a calculation template to calculate the ratio of the known real area of ​​the calibration plate to its corresponding pixel area in the complete kelp image, and multiply this ratio by the kelp phenotypic data represented by the pixels to obtain the real kelp phenotypic data.

[0015] Furthermore, in step 2, the SIFT algorithm is used to detect feature points in each kelp local image; for each feature point, the SIFT algorithm further calculates the descriptor of the local area where it is located; and the Euclidean distance is used to calculate the similarity between the descriptors in consecutive kelp local images.

[0016] Furthermore, in step 3,

[0017] The homography moment calculation method is to calculate the randomly selected matching feature points through the RANSAC algorithm to obtain multiple candidate homography matrices, and then select the most appropriate homography matrix through verification.

[0018] Furthermore, in step 4,

[0019] The process of perspective transformation is to multiply the pixel coordinates in each kelp local image by the homography matrix, thereby mapping the kelp local image A to the coordinate system of the kelp local image B, so that the two kelp local images are aligned in the same coordinate system.

[0020] Furthermore, in step 5,

[0021] Image fusion is performed using weighted averaging to eliminate edge seams and overlapping areas.

[0022] Secondly,

[0023] The present invention provides a kelp phenotypic trait measuring device, comprising a base, characterized in that it also comprises an active pressure roller and a driven pressure roller, two lower supports are provided on the base, the active pressure roller is connected between the two lower supports, and a motor is coaxially connected to the active pressure roller; the top of the lower support is connected to the upper support, and the driven pressure roller is connected between the two lower supports;

[0024] A loading platform connected to the base is provided between the active pressing roller and the driven pressing roller, serving as the measurement area for the kelp to be laid flat when entering; a calibration plate is provided on the loading platform;

[0025] Driven by the motor, the active pressing roller and the driven pressing roller perform rolling, flattening the kelp and transferring it away from the loading platform.

[0026] Furthermore, cross beams are connected between the top ends of the upper supports.

[0027] Furthermore, soft elastic structures are provided on both sides of the upper support, one end of the soft elastic structure is connected to the driven roller pressure wheel, and the other end is connected to the loading platform.

[0028] Furthermore, a motor rear support is provided on the base.

[0029] Furthermore, it also includes a camera arranged above the stage for photographing the kelp on the stage, and uploading a plurality of local kelp images to a host computer, which executes the aforementioned kelp phenotypic trait measurement method.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] 1. The method for measuring kelp phenotypic traits of the present invention comprises the following steps: laying a calibration plate and kelp on a stage, taking multiple local images, extracting feature points and calculating descriptors, finding matching feature point pairs, calculating a homography matrix, image registration and splicing, image fusion, extracting kelp sample areas and calculating true phenotypic data; by using the calibration plate as a calculation template, the true value of the kelp phenotypic data can be accurately calculated, thereby achieving accurate measurement of the kelp phenotypic traits.

[0032] Compared with traditional manual measurement methods, this method significantly improves measurement speed and efficiency and reduces labor costs.

[0033] 2. The present invention uses a method and device for measuring phenotypic traits of kelp to achieve automated out-of-water measurement of phenotypic traits of large kelp, and to quickly measure phenotypic traits such as seaweed leaf length, width, and surface area, providing support for the genetic analysis of important economic traits of large seaweed and the breeding of high-quality varieties with complex traits. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a flow chart of the method for measuring phenotypic traits of kelp of the present invention;

[0035] Figure 2 Schematic diagram of the structure of the device for measuring phenotypic traits of kelp of the present invention;

[0036] Figure 3 It is a rear view of the kelp phenotypic trait measuring device of the present invention.

[0037] Figure 4 The figure is a schematic diagram of the properties of kelp calculated based on the measurement method of the present invention, including length, width and surface area.

[0038] Reference numerals:

[0039] 1 is the crossbeam; 2 is the upper support; 3 is the lower support; 4 is the calibration plate; 5 is the loading platform; 6 is the base; 7 is the motor rear support; 8 is the active pressure roller; 81 is the transmission shaft; 9 is the driven pressure roller; 10 is the motor; 11 is the connecting hole. DETAILED DESCRIPTION

[0040] The technical solution of the present invention will be clearly described below in conjunction with the accompanying drawings. Obviously, the described embodiments are not all embodiments of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0041] It should be noted that the terms "center", "up", "down", "horizontal", "left", "right", "front", "back", "lateral", "longitudinal", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description. They do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0042] Example 1

[0043] like Figure 1 As shown, the present invention provides a method for measuring phenotypic traits of kelp, comprising the following steps:

[0044] Step 1: Set the shooting start condition, shooting frequency, and shooting end condition to collect multiple local images of kelp:

[0045] When the proportion of low pixel value areas in the kelp measurement area, that is, the proportion of the area where the kelp sample is located, exceeds the set threshold, the shooting process is started; if the set threshold is equal to 20%, it serves as the starting condition for shooting;

[0046] The camera continuously shoots at a set frame rate to ensure sufficient image overlap for stitching; for example, the set frame rate is equal to 5 frames per second as the shooting frequency;

[0047] When the high pixel value area on the screen, that is, the area where the kelp background is located, exceeds the set threshold, such as 90%;

[0048] If the structural similarity score between consecutive images exceeds a set threshold, such as 0.98, and lasts for a certain period of time, such as 3 seconds, the shooting is stopped.

[0049] Step 2: Feature point extraction and matching:

[0050] First, the feature points of the kelp are extracted from each local kelp image. Specifically, the SIFT algorithm is used to detect the feature points in each local kelp image.

[0051] SIFT (Scale-Invariant Feature Transform) algorithm: The SIFT algorithm uses a multi-scale Gaussian difference method to extract significant feature points in the local image of kelp that are invariant to scale and rotation.

[0052] For each feature point, the SIFT algorithm further calculates the descriptor of the local area where it is located. The descriptor is generated based on the pixel gradient direction and can uniquely identify the feature point.

[0053] Calculate the similarity of descriptors between consecutive kelp local images, such as using Euclidean distance to calculate the similarity between descriptors in the image; and find matching feature point pairs between consecutive kelp local images based on the similarity of the descriptors.

[0054] In a specific embodiment, the feature point extraction and matching code is as follows:

[0055] stitcher=Stitcher(detector="sift",confidence_threshold=0.1)

[0056] The Stitcher class extracts and matches feature points by setting the SIFT algorithm as the feature extractor. The calculation and matching process of feature point descriptors are encapsulated within this class.

[0057] Step 3. Calculate the homography matrix:

[0058] After extracting the matching feature point pairs, the homography matrix between the two consecutive kelp partial images is calculated. The homography matrix can describe the perspective transformation relationship between the two consecutive kelp partial images.

[0059] The homography moment calculation method is based on feature point matching, and the RANSAC algorithm is used to eliminate incorrectly extracted feature points to ensure the accuracy of the calculation results.

[0060] The RANSAC (Random Sampling Consensus) algorithm obtains multiple candidate homography matrices by calculating randomly selected feature points, and selects the most appropriate homography matrix by verifying whether most feature points conform to the homography transformation.

[0061] In a specific embodiment, the code for calculating the homography matrix is ​​as follows:

[0062] panorama=stitcher.stitch(image_paths)

[0063] In the above code, the stitch method call will calculate the homography matrix based on the extracted matching feature points and perform image registration.

[0064] Step 4: Image registration and transformation:

[0065] After the homography matrix is ​​calculated, the local kelp image can be perspective transformed so that two consecutive local kelp images are aligned in the same coordinate system.

[0066] The perspective transformation process is achieved by multiplying the pixel coordinates in each kelp local image with the homography matrix, thereby mapping the kelp local image A to the coordinate system of the kelp local image B, so that the two kelp local images are aligned in the same coordinate system, achieving the kelp image registration effect and realizing the preliminary splicing of the complete kelp image.

[0067] Step 5: Image fusion:

[0068] The stitched complete kelp image may have edge seams and overlapping areas. Image fusion is performed through weighted averaging or other fusion technologies to eliminate the edge seams and overlapping areas and generate a smooth panoramic complete kelp image.

[0069] By using the weighted average method, a smooth transition can be performed based on the pixel values ​​of the overlapping areas of the two kelp local images to avoid the appearance of stitching lines.

[0070] Step 6: Phenotypic data extraction:

[0071] Through segmentation algorithms, such as the SAM method based on semantic segmentation, the kelp sample area in the spliced ​​complete kelp image is extracted and the number of pixels it occupies is calculated.

[0072] The known real area of ​​the calibration plate is converted to the actual kelp phenotypic number using the following formula by comparing it with the corresponding pixel area in the image:

[0073] S=S'×A / A',

[0074] Among them, S′ is the kelp phenotypic data expressed in pixels, A′ is the calibration plate area expressed in pixels, A is the actual area of ​​the calibration plate, and S is the actual kelp phenotypic data.

[0075] Example 2

[0076] Combine Figure 1 、 Figure 2 The present invention provides a kelp phenotypic trait measurement device for rapidly measuring kelp phenotypic traits such as length, width, thickness, and surface area. This device serves as a fast, convenient, and accurate measurement method during kelp breeding. The device utilizes a roller-type structure to maintain kelp transport stability.

[0077] The maximum size of the kelp to be measured is 4m*0.6m, the roller structure transmission speed is 0.3m / s, the measurement accuracy of the kelp length and width is ±5mm, the measurement accuracy of the kelp fresh weight is ±10g, and the measurement accuracy of the kelp thickness is ±0.5mm.

[0078] The algae phenotypic trait measuring device comprises a base 6 for fixing and supporting all components.

[0079] A stage 5 is provided on the base 6, which serves as a measurement area that is laid out flat when the kelp enters, and the kelp is photographed and spliced ​​in this measurement area; as a measurement platform, the stage 5 needs to ensure that the background color of the measurement platform forms a significant contrast with the sample for subsequent image processing, such as the measurement platform is white.

[0080] Lower supports 3 are symmetrically arranged on the left and right sides of the base 6, and active pressing rollers 8 are connected between the lower supports 3. The right end of the transmission shaft 81 of the active pressing roller 8 is coaxially connected to a motor (not shown in the figure). Driven by the motor, the active pressing roller 8 rotates to drive the kelp forward.

[0081] The right end of the transmission shaft 81 of the active pressure roller 8 is provided with a motor rear support 7 for supporting and sharing the weight of the motor 10, thereby increasing the stability and reliability of the entire device structure.

[0082] The top of the lower support 3 is connected to the upper support 2, and the upper support 2 is connected with a driven pressure roller 9; the active pressure roller 8 and the driven pressure roller 9 perform rolling to flatten the kelp and transport it away from the loading platform 5.

[0083] A portion of the top of the loading platform 5 is located between the active pressure roller 8 and the driven pressure roller 9 .

[0084] A calibration plate 4 is provided on the top of the stage 5, on which grids or lines of known sizes are printed. These grids or lines serve as reference objects to determine the actual size of the kelp during the image processing process, so as to calculate the actual length, width and projected surface area of ​​the kelp.

[0085] The top ends of the upper supports 2 are connected with crossbeams 1 to increase the stability and reliability of the entire device.

[0086] Both the upper and lower struts 2 and 3 are provided with connection holes 11 for inserting connecting posts. A flexible elastic structure (not shown) is connected between the connecting posts of the upper and lower struts 2 and 3. The driven roller 9, connected to the flexible elastic structure through the connection holes 11 on either side, provides pressure to press the kelp. This rolling process flattens the kelp while transporting it for measurement.

[0087] In a specific embodiment, the soft elastic structure adopts a rubber band to tighten the active pressure roller 8 and the driven pressure roller 9.

[0088] In a specific embodiment, the soft elastic structure adopts a soft spring.

[0089] A camera (not shown in the figure) is provided above the loading platform 5 for photographing the kelp on the loading platform 5. As the kelp is transported, the camera continuously acquires images and uploads the photographed kelp images to the host computer connected to the camera for subsequent image processing and analysis.

[0090] First, the kelp is spread flat on the stage 5, the camera captures the image of the kelp on the stage 5, and multiple photos are taken to cover the entire kelp area. The upper computer uses the kelp phenotypic trait measurement method of Example 1 to merge these photos into a complete kelp image to calculate the length, width and projected surface area of ​​the kelp.

[0091] Experimental process:

[0092] 1. Place the root of the kelp steadily on the measuring area of ​​the loading platform 5, then start the motor 10 to transport the kelp. The active pressing roller 8 is powered by the motor 10 and its rotation speed is 0.3 m / s.

[0093] 2. As motor 10 rotates, the kelp is drawn in and gradually transported away from stage 5. Simultaneously, the camera captures an image of the kelp sample in the measurement area. The host computer determines whether the kelp appears on the screen by looking at low-pixel value areas in the measurement area (the measurement platform is white and has higher pixel values, while the kelp is brown and has lower pixel values). When the non-white pixel area on the screen exceeds 20%, kelp is determined to be present, and the camera begins recording at a rate of 5 times per second. Recording stops when the proportion of white pixels on the screen exceeds 90%, and the image pixels remain unchanged for 3 consecutive seconds (a structural similarity index (SSIM) score of two consecutive images is greater than 0.98, indicating that the image has not changed).

[0094] 3. The driven pressure roller 9 uses the soft elastic structures on both sides of the upper support 2 to provide pressure to press the kelp. After the kelp has completely passed through the pressure roller, the motor 10 is controlled to stop rotating. After the kelp has been transported, the image splicing begins.

[0095] 4. The host computer extracts key frames in real time from the video stream uploaded by the camera, and uses the measurement method of Example 1 to splice the images of various parts of the recorded images to obtain a spliced ​​complete kelp picture;

[0096] The basic phenotype of the splicing results was determined, and the real measurement results were obtained according to the calibration, and the length, width and surface area of ​​the kelp were calculated.

[0097] In a specific experiment,

[0098] Experimental kelp samples:

[0099] variety source Sampling date "Haijia No. 1" kelp Fujian Yijia Kelp Seedling Industry Co., Ltd. October 5, 2024

[0100] (1) Manual measurement:

[0101] The length of the "Haijia No. 1" kelp is 36.3 cm, the width is 127.1 cm, and the surface area is 410.8 cm. 2 .

[0102] (2) Using the measuring device and measuring method of the present invention:

[0103] like Figure 4 As shown, the length of the "Haijia No. 1" kelp is 37 cm, the width is 130 cm, and the surface area is 416.59 cm2. For details, see the status bar module at the bottom of the software window developed based on the measurement method of the present invention. The module is responsible for displaying the properties of the kelp calculated after a single measurement, including length, width, and surface area.

[0104] Compared with manual measurement, the error in length is 1.93%, the error in width is 2.28%, and the error in surface area measurement is 1.41%.

[0105] The measurement method of the present invention was used to measure a plurality of kelp samples, and the measurement results are shown in Table 1.

[0106] Table 1 Comparison between roller-type kelp measurement data and manual measurement data

[0107]

[0108]

[0109]

[0110] Since there are currently no commercially available devices for measuring the phenotypic characteristics of large algae, manual measurements were used as the true values ​​to evaluate machine measurement errors. The roller-based kelp phenotyping device achieved a length error of 1.17% and a width error of 4.30%, demonstrating superior length measurement performance. The larger width error is attributed to the roller flattening of the kelp during transmission, resulting in a certain degree of oscillation error. This results in errors in the image stitching results, reducing the accuracy of width measurements.

[0111] The above technical features constitute the best embodiment of the present invention, which has strong adaptability and best implementation effect. Non-essential technical features can be added or removed according to actual needs to meet the needs of different situations.

[0112] Finally, it should be noted that the above content is only used to illustrate the technical solution of the present invention, rather than to limit the scope of protection of the present invention. Simple modifications or equivalent substitutions of the technical solution of the present invention by ordinary technicians in this field do not deviate from the essence and scope of the technical solution of the present invention.

Claims

1. A method for measuring phenotypic traits of kelp, characterized in that: The steps include: Step 1: Place the calibration plate and kelp flat on the stage, flatten the kelp, and take multiple partial images of the kelp; Step 2: Extract the feature points of kelp from each local image of kelp; for each feature point, calculate the descriptor of the local area where it is located; Calculate the similarity of descriptors between consecutive kelp local images; find matching feature point pairs between consecutive kelp local images based on the similarity of descriptors; Step 3: After finding the matching feature point pairs, calculate the homography matrix between the two consecutive kelp local images; Step 4: After the homography matrix is ​​calculated, the perspective transformation is performed on the local kelp image so that two consecutive local kelp images are aligned in the same coordinate system to achieve the kelp image registration effect and obtain a preliminary spliced ​​complete kelp image; Step 5: Perform image fusion on the preliminarily stitched complete kelp images to generate a smooth complete kelp image; Step 6: Using a segmentation algorithm, extract the kelp sample area from the smooth complete kelp image and calculate the number of pixels it occupies; Using the calibration plate as a calculation template, the ratio of the known real area of ​​the calibration plate to its corresponding pixel area in the complete kelp image is calculated. This ratio is multiplied by the kelp phenotypic data represented by the pixels to obtain the real kelp phenotypic data.

2. The method for measuring phenotypic traits of kelp according to claim 1, wherein In step 2, the SIFT algorithm is used to detect feature points in each kelp local image; for each feature point, the SIFT algorithm further calculates the descriptor of the local area where it is located; and the Euclidean distance is used to calculate the similarity between the intermediate descriptors of consecutive kelp local images.

3. The method for measuring phenotypic traits of kelp according to claim 1, wherein In step 3, the homography moment calculation method is to calculate the randomly selected matching feature points through the RANSAC algorithm to obtain multiple candidate homography matrices, and then select the most appropriate homography matrix through verification.

4. The method for measuring phenotypic traits of kelp according to claim 1, wherein In step 4, the perspective transformation process is to multiply the pixel coordinates in each kelp partial image by the homography matrix, thereby mapping the kelp partial image A to the coordinate system of the kelp partial image B, so that the two kelp partial images are aligned in the same coordinate system.

5. The method for measuring phenotypic traits of kelp according to claim 1, wherein In step 5, image fusion is performed using a weighted average method to eliminate edge seams and overlapping areas.

6. A device for measuring phenotypic traits of kelp, comprising a base, characterized in that: It also includes an active pressure roller and a driven pressure roller. Two lower supports are provided on the base. The active pressure roller is connected between the two lower supports. A motor is coaxially connected to the active pressure roller. The top of the lower support is connected to the upper support. The driven pressure roller is connected between the two upper supports. A loading platform connected to the base is provided between the active pressing roller and the driven pressing roller, serving as the measurement area for the kelp to be laid flat when entering; a calibration plate is provided on the loading platform; Driven by the motor, the active pressing roller and the driven pressing roller perform rolling, flattening the kelp and transferring it away from the loading platform.

7. The device for measuring phenotypic properties of kelp according to claim 6, wherein: The top ends of the upper supports are connected with cross beams.

8. The device for measuring phenotypic properties of kelp according to claim 6, wherein: The upper support and the lower support are connected with a soft elastic structure.

9. The device for measuring phenotypic properties of kelp according to claim 6, wherein: A motor rear support is provided on the base.

10. The device for measuring phenotypic characteristics of kelp according to claim 6, characterized in that: It also includes a camera arranged above the stage for photographing the kelp on the stage, and uploading a plurality of local kelp images after photographing to a host computer, and the host computer executes the kelp phenotypic trait measurement method according to any one of claims 1-5.

Images