A three-dimensional reconstruction method for quantifying curly trait of chinese cabbage
By using three-dimensional reconstruction and multiple linear regression analysis, the problem of lack of objective evaluation in the study of the curling trait of Chinese cabbage was solved, and efficient and accurate quantification of the curling trait was achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-12-20
- Publication Date
- 2026-03-17
AI Technical Summary
Current research on the curling characteristics of Chinese cabbage lacks objective evaluation standards. Traditional methods rely on human judgment, resulting in large errors, low efficiency, and high manpower and time consumption.
A three-dimensional reconstruction method was adopted to construct a point cloud model by acquiring a sequence of Chinese cabbage images. After preprocessing, phenotypic shape parameters were extracted, an evaluation standard for the degree of curling was established, and the degree of curling was quantified using multiple linear regression analysis.
This method enables efficient and objective quantification of the curling trait in Chinese cabbage, reduces human error, improves efficiency, and provides accurate evaluation standards.
Smart Images
Figure CN116012528B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of crop phenotypic trait research technology, and in particular to a three-dimensional reconstruction method for quantifying the curling trait of Chinese cabbage. Background Technology
[0002] The three-dimensional morphology of plants is a true reflection of their growth and development, and establishing three-dimensional plant models has always been a research hotspot in botany, computer graphics, and other fields. By utilizing advanced sensor technology and modern information technology to realistically reproduce the morphological structure of crops in a three-dimensional visual manner on a computer, the dynamic processes of plant growth and the plant-environment interaction can be analyzed, simulated, and predicted. Based on multi-view images, the combination of motion reconstruction structure and multi-view stereo vision has been widely used for the three-dimensional reconstruction of plant phenotypes, but research on the three-dimensional model of Chinese cabbage is still under exploration.
[0003] Chinese cabbage originated in northern my country and is the country's second largest vegetable crop, with an annual planting area of approximately 1.8 million hectares, or about 27 million mu. Its edible part is the head (or head bulb). The formation of the head bulb and the curling of the leaves are complex biological processes, as well as morphological responses induced by environmental factors. Leaf curling is a very important trait, directly related to the yield and marketability of Chinese cabbage. However, current research on curling traits is mostly at the genetic level, with relatively few studies focusing on morphology.
[0004] Currently, the assessment of cabbage leaf curling relies heavily on simple visual judgment, depending on subjective perception of whether and to what extent the leaves are curled, without any objective evaluation standards. A small portion of the quantification of curling involves simple two-dimensional manual measurement, which suffers from significant human error, resulting in poor accuracy and low efficiency. Therefore, the current research field lacks accurate, objective, and clear evaluation standards and methods for quantifying cabbage curling. Summary of the Invention
[0005] To overcome the shortcomings of existing technologies, the purpose of this invention is to provide a three-dimensional reconstruction method for quantifying the curling trait of Chinese cabbage. By quantitatively extracting and analyzing the phenotypic parameters of Chinese cabbage, this method achieves efficient, objective, and accurate quantification of the curling trait, thereby solving the problems of inaccurate data, significant crop damage, and high manpower and time consumption caused by traditional methods mentioned in the background.
[0006] To achieve the above objectives, the present invention provides the following solution:
[0007] A three-dimensional reconstruction method for quantifying the curling trait of Chinese cabbage includes:
[0008] Obtain the image sequence of the Chinese cabbage to be tested;
[0009] Construct a point cloud model based on the image sequence;
[0010] The point cloud model is preprocessed to obtain a preprocessed point cloud model;
[0011] Based on the preprocessed point cloud model, phenotypic shape parameters are extracted to obtain extracted data;
[0012] An evaluation standard representing the degree of curling of the Chinese cabbage to be tested is established based on the extracted data.
[0013] Preferably, acquiring the image sequence of the Chinese cabbage to be tested includes:
[0014] The Chinese cabbages to be tested were classified according to different heading types and copulation types;
[0015] Take the same number of leaves from the same position of the classified Chinese cabbage to be tested and place them on an automatic image sequence acquisition platform to obtain the image sequence.
[0016] Preferably, constructing a point cloud model based on the image sequence includes:
[0017] The image sequence is reconstructed based on the motion reconstruction structure algorithm and the multi-view stereo vision algorithm to obtain the point cloud model.
[0018] Preferably, the step of extracting phenotypic shape parameters from the preprocessed point cloud model to obtain extracted data includes:
[0019] Based on the preprocessed point cloud model, the minimum bounding box is created using principal component analysis to obtain the leaf length, leaf width, and leaf height.
[0020] Based on the preprocessed point cloud model, the convex hull algorithm is used to obtain the blade volume parameters;
[0021] Based on the preprocessed point cloud model, the blade area parameters are obtained using the Alpha-shape surface reconstruction algorithm.
[0022] Based on the preprocessed point cloud model, the Gaussian curvature, principal curvature, and average curvature of the point cloud are calculated by selecting different neighborhood radii to represent the different degrees of curling of the leaves of the Chinese cabbage under test.
[0023] Based on the preprocessed point cloud model, the least squares fitting method is used to fit a sphere to obtain the relevant parameters of the fitted sphere.
[0024] The plane fitting coefficients are obtained by using the random sampling consensus algorithm and the minimum singular value decomposition algorithm.
[0025] Projection is used to obtain the projected point cloud of each different fitting surface, and the projection fitting parameters are obtained by comparing the projected point cloud with the original area.
[0026] Preferably, an evaluation standard representing the degree of curling of the Chinese cabbage to be tested is established based on the extracted data, including:
[0027] A radar map is plotted based on the projection fitting parameters, the leaf length, the leaf width, the leaf height, the leaf volume parameter, the leaf area parameter, the point cloud Gaussian curvature, the principal curvature, and the average curvature.
[0028] Based on different heading and clasping types, the raw data in the radar chart are classified, and multiple linear regression analysis is used to compare the leaf length and leaf width as typical values with the true values. The correlation between each parameter and different curling types is analyzed, and an equation is established with leaf curling as the dependent variable and each parameter as the independent variable.
[0029] The evaluation criteria representing the degree of curling of the Chinese cabbage to be tested are determined based on the equation.
[0030] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:
[0031] This invention provides a three-dimensional reconstruction method for quantifying the curling trait of Chinese cabbage, comprising: acquiring an image sequence of Chinese cabbage to be tested; constructing a point cloud model based on the image sequence; preprocessing the point cloud model to obtain a preprocessed point cloud model; extracting phenotypic shape parameters based on the preprocessed point cloud model to obtain extracted data; and establishing an evaluation standard representing the degree of curling of the Chinese cabbage to be tested based on the extracted data. This invention achieves efficient, objective, and accurate quantification of the curling trait of Chinese cabbage through quantitative extraction and analysis of phenotypic parameters, thus solving the problems of inaccurate data, significant crop damage, and high manpower and time consumption caused by traditional methods. Attached Figure Description
[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0033] Figure 1 This is a schematic diagram of the method provided in an embodiment of the present invention;
[0034] Figure 2 This is a flowchart provided for an embodiment of the present invention. Detailed Implementation
[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0036] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0037] The terms "first," "second," "third," and "fourth," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, including a series of steps, processes, methods, etc., is not limited to the steps listed, but may optionally include steps not listed, or may optionally include other steps inherent to these processes, methods, products, or devices.
[0038] The purpose of this invention is to provide a three-dimensional reconstruction method for quantifying the curling trait of Chinese cabbage, which can solve the problems of inaccurate data, significant crop damage, and high manpower and time consumption caused by traditional methods.
[0039] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0040] Figure 1 This is a schematic diagram of the method provided in an embodiment of the present invention, such as... Figure 1 As shown, this invention provides a three-dimensional reconstruction method for quantifying the curling trait of Chinese cabbage, comprising:
[0041] Step 100: Obtain the image sequence of the Chinese cabbage to be tested;
[0042] Step 200: Construct a point cloud model based on the image sequence;
[0043] Step 300: Preprocess the point cloud model to obtain a preprocessed point cloud model;
[0044] Step 400: Extract phenotypic shape parameters based on the preprocessed point cloud model to obtain extracted data;
[0045] Step 500: Establish an evaluation standard to represent the degree of curling of the Chinese cabbage to be tested based on the extracted data.
[0046] Preferably, acquiring the image sequence of the Chinese cabbage to be tested includes:
[0047] The Chinese cabbages to be tested were classified according to different heading types and copulation types;
[0048] Take the same number of leaves from the same position of the classified Chinese cabbage to be tested and place them on an automatic image sequence acquisition platform to obtain the image sequence.
[0049] Specifically, such as Figure 2 As shown, this embodiment obtains image sequences by capturing images and extracting video frames.
[0050] Furthermore, in order to control extraneous variables, the Chinese cabbages were first classified according to different heading types and clasping types. The same number of leaves from the same position were then placed on an automatic image sequence acquisition platform to automatically acquire image sequences. The acquired image data was then used to reconstruct a point cloud model using the SFM (Structure of Motion) and MVS (Multi-View Stereo Vision) algorithms.
[0051] Preferably, constructing a point cloud model based on the image sequence includes:
[0052] The image sequence is reconstructed based on the motion reconstruction structure algorithm and the multi-view stereo vision algorithm to obtain the point cloud model.
[0053] Specifically, in this embodiment, after acquiring the point cloud model, the point cloud model is preprocessed. Since the growth morphology of Chinese cabbage leaves is relatively uniform compared to other crops, this invention uses principal component analysis to perform rigid transformations such as rotation and translation on the point cloud to unify its orientation. After unifying the orientation, additional processing of uniform magnitude is applied to the point cloud model to address errors caused by different coordinate systems built at different locations. After these two steps, point cloud filtering, point cloud downsampling, and point cloud segmentation are performed as preprocessing operations.
[0054] Preferably, the step of extracting phenotypic shape parameters from the preprocessed point cloud model to obtain extracted data includes:
[0055] Based on the preprocessed point cloud model, the minimum bounding box is created using principal component analysis to obtain the leaf length, leaf width, and leaf height.
[0056] Based on the preprocessed point cloud model, the convex hull algorithm is used to obtain the blade volume parameters;
[0057] Based on the preprocessed point cloud model, the blade area parameters are obtained using the Alpha-shape surface reconstruction algorithm.
[0058] Based on the preprocessed point cloud model, the Gaussian curvature, principal curvature, and average curvature of the point cloud are calculated by selecting different neighborhood radii to represent the different degrees of curling of the leaves of the Chinese cabbage under test.
[0059] Based on the preprocessed point cloud model, the least squares fitting method is used to fit a sphere to obtain the relevant parameters of the fitted sphere.
[0060] The plane fitting coefficients are obtained by using the random sampling consensus algorithm and the minimum singular value decomposition algorithm.
[0061] Projection is used to obtain the projected point cloud of each different fitting surface, and the projection fitting parameters are obtained by comparing the projected point cloud with the original area.
[0062] Preferably, an evaluation standard representing the degree of curling of the Chinese cabbage to be tested is established based on the extracted data, including:
[0063] A radar map is plotted based on the projection fitting parameters, the leaf length, the leaf width, the leaf height, the leaf volume parameter, the leaf area parameter, the point cloud Gaussian curvature, the principal curvature, and the average curvature.
[0064] Based on different heading and clasping types, the raw data in the radar chart are classified, and multiple linear regression analysis is used to compare the leaf length and leaf width as typical values with the true values. The correlation between each parameter and different curling types is analyzed, and an equation is established with leaf curling as the dependent variable and each parameter as the independent variable.
[0065] The evaluation criteria representing the degree of curling of the Chinese cabbage to be tested are determined based on the equation.
[0066] Furthermore, in this embodiment, after obtaining the preprocessed point cloud model, phenotypic shape parameters are extracted. The specific technical process is as follows:
[0067] Based on the point cloud model data obtained after 3D reconstruction and preprocessing, the following parameters are extracted in sequence.
[0068] a) Use principal component analysis to create the minimum bounding box and obtain the leaf length, leaf width, and leaf height.
[0069] b): The convex hull algorithm is used to obtain the blade volume parameters.
[0070] c): The blade area parameters are obtained by using the Alpha-shape surface reconstruction algorithm.
[0071] d): Select different neighborhood radii and calculate the Gaussian curvature, principal curvature and mean curvature of the point cloud to represent different degrees of curling of cabbage leaves.
[0072] e): Use the least squares fitting method to fit the sphere and obtain the relevant parameters of the fitted sphere.
[0073] f): Use the random sampling consensus algorithm and the minimum singular value decomposition algorithm to perform plane fitting and obtain the fitting plane coefficients.
[0074] g): Using projection, obtain the projected point cloud of each different fitted surface, compare it with the original area, and use it as an evaluation index for curling.
[0075] Finally, after proposing the above parameters, the projected areas on different fitting surfaces (fitted plane, fitted curved surface, and fitted sphere) are compared with the original areas to obtain fitting parameters. Then, combined with the extracted leaf length, leaf width, leaf height, leaf area, leaf volume, and curvature, a radar chart is drawn. Finally, the original data are classified according to different calyx types and different spherical shapes. Multiple linear regression analysis is used, with leaf length and leaf width parameters as typical examples, compared with the true values to analyze the correlation between each parameter and different curling types. With leaf curling as the dependent variable and each parameter as the independent variable, an equation is established. Finally, based on these parameters, an evaluation standard that can represent the degree of curling in Chinese cabbage is established.
[0076] This invention applies three-dimensional reconstruction methods, three-dimensional point cloud model processing methods, and multivariate statistical data processing methods to extract parameters from three-dimensional modeling of Chinese cabbage, thereby quantifying the curling trait and ultimately linking it to biology, providing effective support for improving traits and increasing yield and quality.
[0077] The beneficial effects of this invention are as follows:
[0078] (1) This invention fills the gap in the current research field of Chinese cabbage curling standards. It uses automated and digital methods to solve the problems of large human error, easy damage to crops and low efficiency of traditional methods. It realizes the automatic and non-destructive extraction of parameters and accurately and objectively completes the quantification of Chinese cabbage curling.
[0079] (2) In terms of image data acquisition, the present invention classifies cabbage according to different heading types and clasping methods, selects the same number of cabbage leaves at the same position, and adopts a three-dimensional reconstruction method based on multi-view images. Through an automatic acquisition platform, the cabbage is automatically and efficiently acquired using two methods: direct shooting and video frame extraction, in order to obtain a point cloud model.
[0080] (3) In terms of point cloud preprocessing, the present invention takes the unification of point cloud direction and coordinate magnitude as the first step of preprocessing, eliminates irrelevant variables, and improves the accuracy and precision of point cloud reconstruction.
[0081] (4) In terms of phenotypic parameter extraction, the present invention applies the three-dimensional fitting method used in engineering such as workpieces and bridges to cabbage to extract a variety of phenotypic parameters.
[0082] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.
[0083] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.
Claims
1. A three-dimensional reconstruction method for quantifying the curly trait of Brassica rapa, characterized by, The method comprises the following steps: acquiring an image sequence of the to-be-tested Chinese cabbage; constructing a point cloud model according to the image sequence; preprocessing the point cloud model to obtain a preprocessed point cloud model; extracting a phenotype shape parameter according to the preprocessed point cloud model to obtain extraction data; establishing an evaluation standard representing the curling degree of the to-be-tested Chinese cabbage according to the extraction data; the extraction of the phenotype shape parameter according to the preprocessed point cloud model to obtain the extraction data comprises: based on the preprocessed point cloud model, using principal component analysis to create a minimum bounding box to obtain leaf length, leaf width and leaf height; based on the preprocessed point cloud model, using a convex hull algorithm to obtain a leaf volume parameter; based on the preprocessed point cloud model, using an Alpha-shape surface reconstruction algorithm to obtain a leaf area parameter; based on the preprocessed point cloud model, selecting different neighborhood radii to calculate point cloud Gaussian curvature, principal curvature and average curvature to represent different curling degrees of the leaves of the to-be-tested Chinese cabbage; based on the preprocessed point cloud model, using a least squares fitting method for spherical fitting to obtain related parameters of the fitted sphere; using a random sample consensus algorithm and a minimum singular value decomposition algorithm for plane fitting to obtain fitted plane coefficients; using projection to obtain projection point clouds of each different fitted surface, and obtaining projection fitting parameters according to the projection point clouds and the original area; the establishment of the evaluation standard representing the curling degree of the to-be-tested Chinese cabbage according to the extraction data comprises: drawing a radar chart according to the projection fitting parameters, the leaf length, the leaf width, the leaf height, the leaf volume parameter, the leaf area parameter, the point cloud Gaussian curvature, the principal curvature and the average curvature; classifying the original data in the radar chart according to different balling types and clasping types, and using a multivariate linear regression analysis method to take the leaf length and the leaf width as typical values, compare them with true values, analyze the correlation of each parameter with different curling types, and establish an equation with leaf curling as the dependent variable and each parameter as the independent variable; determining the evaluation standard representing the curling degree of the to-be-tested Chinese cabbage according to the equation.
2. The method for three-dimensional reconstruction of the quantitative trait of the curled trait of Chinese cabbage according to claim 1, characterized in that, the acquisition of the image sequence of the to-be-tested Chinese cabbage comprises: classifying the to-be-tested Chinese cabbage into different balling types and clasping types; placing the same number of leaves of the classified to-be-tested Chinese cabbage at the same position on an image sequence automatic acquisition platform to acquire the image sequence.
3. The method of three-dimensional reconstruction for quantifying the curled trait of Chinese cabbage according to claim 1, wherein, the construction of the point cloud model according to the image sequence comprises: based on a motion recovery structure algorithm and a multi-view stereo vision algorithm, reconstructing the image sequence to obtain the point cloud model.
Citation Information
Patent Citations
Crop root and fruit phenotype parameter extraction and root and fruit phenotype discrimination method and system
CN111291686A