Point cloud collection method for green plant time sequence model

Abstract

The invention discloses a point cloud collection method for a green plant time sequence model and belongs to the field of plant three-dimensional morphological phenotype measuring technologies. The method aims to solve the problems that during construction of a plant three-dimensional model, the stem of a target plant is slim, leaves of the target plant are small and soft, common methods generallyrequire plant rotation and sensor fixation, stem and leaf shaking of the plant are easily caused, and consequently the reconstruction quality of the three-dimensional model is low; and meanwhile, anexisting SfM algorithm has the defects of inconsistent coordinate systems, poor repeatability and low reliability in the plant three-dimensional morphological phenotype measuring field. The method canadapt to the image collection mode of plant fixation and sensor motion and can effectively avoid the defects, caused by sensor motion, that sensor position parameters are deficient, precision cannotbe guaranteed, and image noise is excessively high; and the method is a plant three-dimensional point cloud acquisition method meeting the requirement of the plant time sequence model for coordinate system consistency.

Description

technical field [0001] The invention belongs to the technical field of plant three-dimensional morphological phenotype measurement, and in particular relates to a point cloud collection method for a time series model of green plants. Background technique [0002] Plant phenotype refers to a series of physical, physiological and biochemical characteristics and traits that reflect the structure, composition, development process and results of plants; the study of plant phenotypes can guide the study of plant genotypes, and is of great importance to the development of plant genomics and mastering the laws of plant development. Significance. [0003] By measuring the phenotypic parameters of plants and analyzing the changes of various parameters over time, the dynamic growth model of plants can be constructed, that is, the time series model of plants, which is used to visually display the deduction of roots, stems, leaves, flowers and fruits of plants over time. Changes in appe...

AI Technical Summary

Problems solved by technology

[0016] When constructing a 3D model of a plant, the target plant has slender stems and small and soft leaves. Commonly used methods generally require the plants to be rotated and the sensors fixed, which can easily cause the stems and leaves of such plants to vibrate, thus affecting the reconstruction quality of the 3D model. At the same time, the existing SfM algorithm has the problems of inconsistent coordinate system, poor repeatability, and low reliability in the field of plant three-dimensional morphological phenotype measurement. The present invention provides a point cloud acquisition method for green plant time series models, which can adapt to The image acquisition method of fixed plants and sensor movement can effectively avoid the shortcomings of missing sensor position parameters, inability to guarantee accuracy, and too much image noise due to sensor movement.

Images