Method and device for quantifying three-dimensional characteristics of plant root system in colloid culture

Abstract

The invention relates to the technical field of plant phenotype image processing, and particularly discloses a colloid culture plant root system three-dimensional characteristic quantification method and device, which are used for solving the problem that rapid and lossless three-dimensional observation and accurate analysis on a plant root system are difficult to realize by the existing traditional root system measurement tool and measurement method. The method comprises the following steps: culturing a plant root system under a colloid medium, acquiring a plant root system image, pre-processing the plant root system image, performing three-dimensional reconstruction on the plant root system based on SFM and NeRF algorithms, automatically extracting plant root system characteristics based on a three-dimensional reconstruction model, and integrating an optical calibration tank device into an imaging system. The optical refraction of the curved surface of the glass culture bottle can be effectively corrected, a three-dimensional point cloud is reconstructed in combination with a high-precision three-dimensional reconstruction algorithm, lossless in-situ measurement is realized, the dynamic growth process of a root system is captured, and the morphology and distribution characteristics of the root system are analyzed.

Description

Technical Field

[0001] The present invention relates to the technical field of plant phenotype image processing, and more specifically to a method and device for quantifying three-dimensional characteristics of plant roots cultured in colloid. Background Art

[0002] The spatial layout of roots in the soil or growth medium, namely the root configuration, has a profound impact on the growth efficiency and adaptability of plants. The difference in this configuration can significantly regulate the plant's absorption efficiency and distribution pattern of resources (such as water and nutrients), and at the same time affect its ability to cope with environmental stress (such as drought, pests and diseases). However, the particularity of its growth environment and the complexity of its configuration make it difficult for traditional root measurement tools and measurement methods to achieve rapid, non-destructive three-dimensional observation and accurate analysis of plant roots.

[0003] In recent years, the rapid progress of plant phenomics has greatly promoted botanical research, especially the study of plant root morphology based on image analysis technology. It not only realizes high-throughput data processing, ensures the accuracy and non-destructiveness of the research, but also is easy to integrate and expand with other technical platforms. However, revealing the genetic basis and developmental mechanism of root architecture faces many challenges. It is necessary to adopt an interdisciplinary and diversified research strategy to closely combine field observations, greenhouse controlled experiments and laboratory precision analysis. Although field research can directly reflect the actual growth of plants in natural environments, its inherent limitations, such as the destructiveness of root excavation and the uncontrollable and variable environmental factors, seriously restrict the acquisition of accurate data on the genetic characteristics and developmental characteristics of root architecture. To overcome this problem, indirect but efficient methods, such as soil core samples, root-crown ratios or root observations under controlled growth conditions, are used to indirectly infer and predict root structure information. At the same time, in order to obtain the root architecture of plants in situ during growth, a series of high-precision hardware equipment such as three-dimensional laser scanners, magnetic resonance imaging (MRI) and computed tomography (CT) are applied to the acquisition of plant root phenotypes. Although these technologies theoretically provide a high observation depth and accuracy, they are still facing many challenges in large-scale and widespread application due to factors such as equipment cost, operation complexity and data processing efficiency of current technologies. For example, Chinese patent publication number CN101658107B discloses a method for in-situ dynamic observation and measurement of plant roots: a three-dimensional laser scanner is used to scan the root system, and in-situ observation and image capture of the three-dimensional configuration of the root system are performed. However, this requires a lot of scanning time and can only reconstruct a pseudo three-dimensional root system structure. For example, Chinese patent publication number CN111693551B proposes a device and method for non-destructive measurement of three-dimensional traits of rice plants and roots. CT is used to scan the rice root system, and the root system tomographic image is obtained and then reconstructed through an algorithm. Although this method can reconstruct a visualized three-dimensional model of the rice root system, it requires the use of high-precision CT equipment, and there are deficiencies in equipment cost and reconstruction efficiency.Based on obtaining root images from different perspectives for three-dimensional reconstruction, the dependence on equipment is effectively reduced. For example, the Chinese patent with publication number CN118505784A discloses a method for three-dimensional reconstruction and phenotypic feature extraction of hydroponic plant roots. The three-dimensional reconstruction of rice roots under hydroponic conditions was performed using multi-perspective images, and a series of three-dimensional phenotypic characteristic parameters of hydroponic plant roots were extracted. However, the shaking of water flow inevitably affects the in-situ observation of plant roots. A gellan gum growth system with optical clarity has been introduced to promote non-invasive two-dimensional and three-dimensional imaging studies of plant roots. Among them, Leon V. Kochian et al. (Leon V. Kochian, 2011) developed a new imaging and software platform for high-throughput phenotypic analysis of three-dimensional root traits during seedling development, and discussed its great potential in high-throughput root phenotyping and new trait discovery of three-dimensional image sets. However, some topological traits of the root system are still difficult to quantify.

[0004] In order to solve the above problems, a technical solution is now provided. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method and device for quantitatively determining the three-dimensional characteristics of plant roots cultured in colloid, and provides a method for rapid in-situ image acquisition, three-dimensional reconstruction and trait extraction of plant roots. This method can be used to quickly reconstruct the plant root system in three dimensions, and combined with an automated trait extraction algorithm, the root configuration traits can be quantitatively extracted to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for quantitatively determining three-dimensional characteristics of plant roots cultured in colloid, comprising the following steps:

[0008] Step 1, cultivating plant roots in a colloidal medium;

[0009] Step 2: After the plants have grown for 10 days, the plants to be tested are placed on an electric turntable in the optical calibration tank, the electric turntable is controlled to rotate by a controller, and a Raspberry Pi camera fixed on an imaging device is used to take an image of the plants after one rotation, so as to obtain multi-view images of the plants;

[0010] Step 3, pre-processing the multi-view images, cropping the images, and obtaining the target root area images;

[0011] Step 4: Sparsely reconstruct and calculate the camera pose based on the SFM algorithm, and reconstruct the three-dimensional model of the plant root system through the NeRF algorithm;

[0012] Step 5, extracting the phenotypic traits of the three-dimensional root system according to the three-dimensional image processing algorithm; the phenotypic traits include the basic morphology and branching traits of the root system; the basic morphological traits of the root system include depth, volume, surface area, projected area and projected perimeter; the branching traits of the root system include the number of branches and branch length.

[0013] As a further scheme of the present invention, in step 1, the plant root system is cultured in a colloidal medium, specifically comprising the following steps: soaking the plant seeds in a 70% ethanol solution for 1 minute, then soaking them in a 15% sodium hypochlorite solution for 15 minutes, washing the seeds with sterile deionized water at least five times to remove the sodium hypochlorite, placing the plant seeds in the dark at 30° C. on a vertically oriented sterile culture dish to germinate, and when the length of the newly emerged radicle reaches about 1 cm, transplanting the sterile seedlings into a transparent growth medium in a glass cylindrical bottle for culture; the growth medium consists of 1 / 2 Kimura B nutrient solution with a pH of 5.8 and 0.2% Phytagel powder.

[0014] As a further solution of the present invention, the glass cylindrical bottle is made of organic glass, with a height of 20 cm and a diameter of 8 cm.

[0015] As a further solution of the present invention, in step 2, the control program of the electric turntable is set to rotate every 6° during the entire 360° rotation process, with an interval of 2s between each rotation, and take pictures through the Raspberry Pi camera, namely, a sequence of 60 silhouette images taken every 6°, which are stored locally.

[0016] As a further solution of the present invention, in step 3, the specific steps of pre-processing the multi-view images include: performing ROI extraction on the RGB image of the plant root system, and the extracted target root area image is the part of the plant root colloid; converting the target root area image into a grayscale image, and using the Otsu method to obtain a binary image of the root system, that is, the image background is black and the root part image is white; using a connected domain algorithm to filter out small connected domains and retain the root part; obtaining the foreground image of the root system image through logical operations, and graying it.

[0017] As a further solution of the present invention, in step 4, the specific steps of reconstructing the three-dimensional model of the plant root system are: using the COLMAP open source framework for feature extraction and matching, camera parameter estimation, dedistortion and sparse reconstruction; using the algorithm in the instant-NGP open source framework to generate a transform.json file from the camera pose, and finally performing three-dimensional reconstruction of the image of the specified path.

[0018] As a further solution of the present invention, in step 5, the root depth is sorted by a sorting algorithm to sort all point clouds in the vertical direction from small to large, and the difference between the last point cloud and the first point cloud in the sorted list is obtained, and the difference between the two in the vertical direction is the root depth; the surface area and volume of the root system are calculated by triangulating the three-dimensional point cloud through three-dimensional convex hull calculation, and then calculating the volume and surface area of ​​the triangular mesh; the root projection area and projection perimeter are obtained by obtaining the coordinate information of the three-dimensional point cloud of the root part in the x direction and the y direction, and the convex hull of the projected point cloud is drawn. , get a convex polygon, the area and perimeter of the convex polygon are the projected area and perimeter of the root system; the number of root branches and branch length, use the Laplace shrinkage algorithm to get the root skeleton, measure the similarity between point clouds according to the Euclidean distance to cluster them, use the shortest path algorithm to find the path between the connection point and the outermost edge points in the branch group, get the longest branch in the branch group, repeat the above process until the branches in each branch group are separated, get all the branches in each root system to count the number of branches, and calculate the Euclidean distance between every two points of the branch as the branch length.

[0019] A colloidally cultured plant root three-dimensional feature quantification device, used for the above-mentioned colloidally cultured plant root three-dimensional feature quantification method, comprises a PLA resin shell, an optical calibration slot, a background light, a glass culture bottle, an electric turntable, a Raspberry Pi camera and a laptop computer; the background light is located directly behind the optical calibration slot, the electric turntable is placed in the optical calibration slot and is connected to a motor outside the slot through a magnetic coupler, the optical calibration slot is fixed on the PLA resin shell, and the Raspberry Pi camera is directly opposite to the optical calibration slot.

[0020] As a further solution of the present invention, the optical calibration groove includes four transparent plates and an acrylic bottom plate. The four transparent plates are relatively fitted, and the left and right sides and the bottom are set to be sealed. A cavity is formed between the four transparent plates, and the top of the cavity is set to be an open structure; the bottom of the transparent plate is sealed and connected to the acrylic bottom plate, and the electric turntable equipped with the driven part of the magnetic coupler is connected to the acrylic bottom plate, and extends to the bottom of the cavity through the bottom opening to be connected to the active part of the magnetic coupler. The four corners of the optical calibration groove are fixed at the bottom of the PLA resin shell, and the optical calibration groove is enclosed in the PLA resin shell.

[0021] The optical correction slot is 20 cm away from the center of the motorized turntable to minimize reconstruction artifacts that may be caused by the perspective geometry of the imaging system. The optical correction slot is rectangular and is used to correct optical refraction on the curved surface of the glass cylinder. Daily image sequences can be captured for each plant root grown in gellan gum, including 60 silhouette images taken and stored every 6° during the entire 360° rotation. The capture time for a single image sequence is 3 minutes, and the image resolution is 60 mm per pixel. Image processing algorithms have been developed to overcome the impact of camera distortion on reconstruction, ensuring that high-quality phenotypic images can be collected and three-dimensionally reconstructed during the growth of plant roots.

[0022] The technical effects and advantages of the method and device for quantifying three-dimensional characteristics of plant roots cultured in a colloid medium are as follows: the method cultivates plant roots in a colloid medium, obtains plant root images, pre-processes plant root images, and reconstructs plant roots in three dimensions based on SFM and NeRF algorithms, and finally automatically extracts plant root characteristics based on the three-dimensional reconstruction model, and then integrates the optical calibration groove device into the imaging system. The method can effectively correct the optical refraction of the curved surface of the glass culture bottle, reconstruct a three-dimensional point cloud in combination with a high-precision three-dimensional reconstruction algorithm, and realize non-destructive in-situ measurement, which is helpful to capture the dynamic growth process of the root system and analyze the root morphology and distribution characteristics. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 The present invention is a schematic diagram of the process of a method for quantitatively determining three-dimensional characteristics of plant roots cultured in colloid.

[0024] Figure 2 This is a schematic diagram of a device for three-dimensional reconstruction of plant roots cultured in colloid according to the present invention.

[0025] Figure 3 This is a reconstruction of the three-dimensional model of the plant root system of the present invention.

[0026] Figure 4 The trait diagram is obtained by extracting the phenotypic traits of the three-dimensional root system using the three-dimensional image processing algorithm of the present invention.

[0027] In the picture: 1. PLA resin shell; 2. Optical calibration slot; 3. Background light; 4. Glass culture bottle; 5. Electric turntable; 6. Raspberry Pi camera; 7. Laptop computer. DETAILED DESCRIPTION

[0028] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solution in the present invention. Obviously, the described technical solution is only a part of the present invention, not all of it. Based on the technical solution in the present invention, all other technical solutions obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0029] Figure 1 FIG. 1 is a flow chart of a method for quantifying three-dimensional characteristics of plant roots cultured in colloids provided in Example 1 of the present invention. Figure 1 As shown, a method for quantifying three-dimensional characteristics of plant roots cultured in a colloid in this embodiment includes the following steps:

[0030] Step 1, cultivating plant roots in a colloidal medium;

[0031] Step 2: After the plant has grown for 10 days, the plant to be tested is placed on the electric turntable 5 in the optical calibration tank 2, the electric turntable 5 is controlled to rotate by the controller, and a Raspberry Pi camera 6 fixed on the imaging device is used to take an image of the plant after one rotation, so as to obtain a multi-view image of the plant;

[0032] Step 3, pre-processing the multi-view images, cropping the images, and obtaining the target root area images;

[0033] Step 4: Sparsely reconstruct and calculate the camera pose based on the SFM algorithm, and reconstruct the three-dimensional model of the plant root system through the NeRF algorithm;

[0034] Step 5, extracting the phenotypic traits of the three-dimensional root system according to the three-dimensional image processing algorithm; the phenotypic traits include the basic morphology and branching traits of the root system; the basic morphological traits of the root system include depth, volume, surface area, projected area and projected perimeter; the branching traits of the root system include the number of branches and branch length.

[0035] Specifically, in step 1, the plant root system is cultured in a colloidal medium, and the specific steps are: soaking the plant seeds in a 70% ethanol solution for 1 minute, then soaking them in a 15% sodium hypochlorite solution for 15 minutes, washing the seeds with sterile deionized water at least five times to remove the sodium hypochlorite, placing the plant seeds in the dark at 30° C. on a vertically oriented sterile culture dish to germinate, and when the length of the newly emerged radicle reaches about 1 cm, transplanting the sterile seedlings into a transparent growth medium in a glass cylindrical bottle for culture; the growth medium consists of 1 / 2 Kimura B nutrient solution with a pH of 5.8 and 0.2% Phytagel powder.

[0036] The glass cylindrical bottle is made of plexiglass, 20cm high and 8cm in diameter.

[0037] In more detail, the growth medium consists of 1 / 2 Kimura B nutrient solution at a pH of 5.8 and 0.2% Phytagel powder, which has been solidified with gellan gum (Sigma-Aldrich Phytagel). The gellan gum growth medium is prepared by dissolving 0.7 grams of Phytagel powder in 0.35 liters of 18-M water and autoclaving, and then the sterile gellan gum solution is mixed with 0.35 L of 1 / 2 Kimura B nutrient solution, which has been adjusted to a pH of 5.8 and filtered sterilized; the 1 / 2 Kimura B nutrient solution contains the following ingredients: (0.091 mM KH2PO4, 0.273 mM MgSO4·7H2O, 0.182 mM (NH4)2SO4, 0.091 mM KNO3, 0.183 mM Ca(NO3)2·4H2O, 0.003 mM H3BO3, 0.0005 mM MnCl2·4H2O, 0.001mM (NH4)6Mo7O24·4H2O, 0.0004mM ZnSO4·7H2O, 0.0002mM CuSO4·5H2O, 0.02mM Fe(III)-EDTA).

[0038] Specifically, in step 2, the control program set for the electric turntable 5 is to rotate every 6° during the entire 360° rotation process, with an interval of 2s between each rotation, and take photos through the Raspberry Pi camera 6, namely, a sequence of 60 silhouette images taken every 6°, which are stored locally.

[0039] Specifically, in step 3, the specific steps of pre-processing the multi-view image include: performing ROI extraction on the RGB image of the plant root system, and the extracted target root area image is the part of the plant root colloid; converting the target root area image into a grayscale image, and using the Otsu method to obtain a binary image of the root system, that is, the image background is black and the root part image is white; using the connected domain algorithm to filter out small connected domains and retain the root part; obtaining the foreground image of the root system image through logical operations, and graying it.

[0040] Specifically, in step 4, the specific steps for reconstructing the three-dimensional model of the plant root system are: using the COLMAP open source framework for feature extraction and matching, camera parameter estimation, dedistortion and sparse reconstruction; using the algorithm in the instant-NGP open source framework to generate a transform.json file from the camera pose, and finally performing three-dimensional reconstruction of the image of the specified path.

[0041] Specifically, in step 5, the root depth is sorted by a sorting algorithm to sort all point clouds in the vertical direction from small to large, and the difference between the last point cloud and the first point cloud in the sorted list is obtained. The vertical difference between the two is the root depth; the surface area and volume of the root system are calculated by triangulating the three-dimensional point cloud through three-dimensional convex hull calculation, and then calculating the volume and surface area of ​​the triangular mesh; the root projection area and projection perimeter are obtained by obtaining the coordinate information of the three-dimensional point cloud of the root part in the x and y directions, and drawing the convex hull of the projected point cloud to obtain a The root system is a convex polygon whose area and perimeter are the projected area and perimeter of the root system. The root system branches and branch length are calculated by using the Laplace shrinkage algorithm to obtain the root system skeleton. The point clouds are clustered according to the Euclidean distance between them. The shortest path algorithm is used to find the path between the connection point and the outermost edge points in the branch group to obtain the longest branch in the branch group. The above process is repeated until all branches in each branch group are separated. The number of branches in each root system is obtained and the Euclidean distance between every two points on the branch is calculated as the branch length.

[0042] A colloidally cultured plant root three-dimensional feature quantification device, used for the above-mentioned colloidally cultured plant root three-dimensional feature quantification method, comprises a PLA resin shell 1, an optical calibration slot 2, a background light 3, a glass culture bottle 4, an electric turntable 5, a Raspberry Pi camera 6 and a laptop computer 7; the background light 3 is located directly behind the optical calibration slot 2, the electric turntable 5 is placed in the optical calibration slot 2 and is connected to a motor outside the slot through a magnetic coupler, the optical calibration slot 2 is fixed on the PLA resin shell 1, and the Raspberry Pi camera 6 is directly opposite to the optical calibration slot 2.

[0043] Specifically, the optical calibration slot 2 includes four transparent plates and an acrylic bottom plate. The four transparent plates are relatively fitted, and the left and right sides and the bottom are set to be sealed. A cavity is formed between the four transparent plates, and the top of the cavity is set to be an open structure; the bottom of the transparent plate is sealed and connected to the acrylic bottom plate, and the electric turntable 5 equipped with the driven part of the magnetic coupler is connected to the acrylic bottom plate, and extends to the bottom of the cavity through the bottom opening to be connected to the active part of the magnetic coupler. The four corners of the optical calibration slot 2 are fixed at the bottom of the PLA resin shell 1, and the optical calibration slot 2 is enclosed in the PLA resin shell 1.

[0044] The optical correction slot is 20 cm away from the center of the motorized turntable to minimize reconstruction artifacts that may be caused by the perspective geometry of the imaging system. The optical correction slot is rectangular and is used to correct the optical refraction of the curved surface of the glass cylinder. Daily image sequences can be captured for each plant root grown in gellan gum, including 60 silhouette images taken and stored every 6° during the entire 360° rotation. The capture time for a single image sequence is 3 minutes, and the image resolution is 60 mm per pixel. Image processing algorithms have been developed to overcome the impact of camera distortion on reconstruction, which can ensure that high-quality phenotypic images can be collected and three-dimensionally reconstructed during the growth of plant roots.

[0045] The embodiment of the present invention cultivates plant roots in a colloidal medium, acquires plant root images, pre-processes plant root images, reconstructs plant roots in three dimensions based on SFM and NeRF algorithms, and finally automatically extracts plant root features based on the three-dimensional reconstruction model. The optical calibration slot 2 device is then integrated into the imaging system. This can effectively correct the optical refraction of the curved surface of the glass culture bottle 4, reconstruct a three-dimensional point cloud using a high-precision three-dimensional reconstruction algorithm, and achieve non-destructive in-situ measurement, which is helpful to capture the dynamic growth process of the root system and analyze the root morphology and distribution characteristics.

[0046] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0047] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for quantifying three-dimensional characteristics of plant roots cultured in colloid, characterized in that: The steps include: Step 1, cultivating plant roots in a colloidal medium; Step 2: After the plant has grown for 10 days, the plant to be tested is placed on the electric turntable (5) in the optical calibration tank (2), the electric turntable (5) is controlled to rotate by the controller, and a Raspberry Pi camera (6) fixed on the imaging device is used to take an image of the plant after one rotation, so as to obtain a multi-view image of the plant; Step 3, pre-processing the multi-view images, cropping the images, and obtaining the target root area images; Step 4: Sparsely reconstruct and calculate the camera pose based on the SFM algorithm, and reconstruct the three-dimensional model of the plant root system through the NeRF algorithm; Step 5, extracting the phenotypic traits of the three-dimensional root system according to the three-dimensional image processing algorithm; the phenotypic traits include the basic morphology and branching traits of the root system; the basic morphological traits of the root system include depth, volume, surface area, projected area and projected perimeter; the branching traits of the root system include the number of branches and branch length.

2. The method for quantifying three-dimensional characteristics of plant roots cultured in colloid according to claim 1, characterized in that: In step 1, the plant root system is cultured in a colloidal medium, specifically, the plant seeds are soaked in a 70% ethanol solution for 1 minute, then soaked in a 15% sodium hypochlorite solution for 15 minutes, the seeds are washed with sterile deionized water at least five times to remove the sodium hypochlorite, the plant seeds are placed in the dark at 30°C on a vertically oriented sterile culture dish to germinate, and when the length of the newly emerged radicle reaches about 1 cm, the sterile seedlings are transplanted into a glass cylindrical bottle and cultured in a transparent growth medium; the growth medium is composed of a pH of 5.8 1 / 2Kimura B nutrient solution and 0.2% Phytagel powder.

3. The method for quantifying three-dimensional characteristics of plant roots cultured in colloid according to claim 1, characterized in that: In step 2, the control program of the electric turntable (5) is set to rotate every 6° during the entire 360° rotation process, with an interval of 2s between each rotation, and take photos through the Raspberry Pi camera, namely, a sequence of 60 silhouette images taken every 6°, which are stored locally.

4. The method for quantifying three-dimensional characteristics of plant roots cultured in colloid according to claim 1, characterized in that: In step 3, the specific steps of pre-processing the multi-view image include: performing ROI extraction on the RGB image of the plant root system, and the extracted target root area image is the colloid part of the plant root system; converting the target root area image into a grayscale image, and using the Otsu method to obtain a binary image of the root system, that is, the image background is black and the root part image is white; using the connected domain algorithm to filter out small connected domains and retain the root part; obtaining the foreground image of the root system image through logical operations, and graying it.

5. The method for quantifying three-dimensional characteristics of plant roots cultured in colloid according to claim 1, characterized in that: In step 4, the specific steps for reconstructing the 3D model of the plant root system are as follows: using the COLMAP open source framework for feature extraction and matching, camera parameter estimation, dedistortion and sparse reconstruction; using the algorithm in the instant-NGP open source framework to generate a transform.json file from the camera pose, and finally performing 3D reconstruction on the image of the specified path.

6. The method for quantifying three-dimensional characteristics of plant roots cultured in colloid according to claim 1, characterized in that: In step 5, the root depth is sorted by a sorting algorithm to sort all point clouds from small to large in the vertical direction, and the difference between the last point cloud and the first point cloud in the sorted list is obtained. The vertical difference between the two is the root depth; the surface area and volume of the root system are calculated by triangulating the three-dimensional point cloud through three-dimensional convex hull calculation, and then calculating the volume and surface area of ​​the triangular mesh; the root projection area and projection perimeter are obtained by obtaining the coordinate information of the three-dimensional point cloud of the root part in the x and y directions, and drawing the convex hull of the projected point cloud to obtain a convex hull. Polygon, the area and perimeter of the convex polygon are the projected area and perimeter of the root system; the number of root branches and the length of the branches, use the Laplace shrinkage algorithm to obtain the root skeleton, cluster them according to the Euclidean distance between the point clouds, and use the shortest path algorithm to find the path between the connection point and the outermost edge points in the branch group to obtain the longest branch in the branch group. Repeat the above process until the branches in each branch group are separated, obtain all the branches in each root system, and count the number of branches. Calculate the Euclidean distance between every two points on the branch, which is the branch length.

7. A colloid-cultured plant root three-dimensional characteristic quantitative device, used to implement a colloid-cultured plant root three-dimensional characteristic quantitative method according to any one of claims 1 to 6, characterized in that: The device comprises a PLA resin shell (1), an optical calibration slot (2), a background light (3), a glass culture bottle (4), an electric turntable (5), a Raspberry Pi camera (6) and a laptop computer (7); the background light (3) is located directly behind the optical calibration slot (2); the electric turntable (5) is placed in the optical calibration slot (2) and is connected to a motor outside the slot through a magnetic coupler; the optical calibration slot (2) is fixed on the PLA resin shell (1); and the Raspberry Pi camera (6) is directly opposite to the optical calibration slot (2).

8. The device for quantifying three-dimensional characteristics of plant roots cultured in colloid according to claim 1, characterized in that: The optical calibration groove (2) comprises four transparent plates and an acrylic bottom plate. The four transparent plates are relatively fitted, and the left and right sides and the bottom are all arranged to be sealed and connected. A cavity is formed between the four transparent plates, and the top of the cavity is arranged to be an open structure. The bottom of the transparent plate is sealed and connected to the acrylic bottom plate. The electric turntable (5) equipped with the driven part of the magnetic coupler is connected to the acrylic bottom plate and extends to the bottom of the cavity through the bottom opening to be connected to the active part of the magnetic coupler. The four corners of the optical calibration groove (2) are fixed to the bottom of the PLA resin shell (1), and the optical calibration groove (2) is surrounded by the PLA resin shell (1).

Images