Research On Three Dimensional Digitization And Architecture Analysis Of Maize Plant And Canopy

Crop colony is the organization system which performs photosynthesis and matter production function.Its morphological structure has important influence on light interception ability,canopy photosynthetic efficiency and crop yields.The morphological characteristics of crop colony has always been the most basic way for people to recognize,analyze and evaluate crops.However,the morphological structure of crop colony has the characteristics of high complexity,poor spatial distribution regularity,high organ surface structure variability,and there are a large number of organs with occlusion,intersection and interaction.Therefore,the morphological structure of crop colony is not a physical process of simple replication.Morphology study on traditional agriculture for crop colony were mainly based on experience experiments of manual measurements,it is difficult to accurately describe the morphological differences caused by crop varieties,planting density,and artificial managements.Therefore,it is of great practical significance to rapidly and accurately model and analyze the morphology and structure of crop colony by integrating the technology and methods of computer graphics,statistics,and agriculture in a digital and visual way.Maize is one of the most important grain crops in China with great potential for increasing yields.At the same time,maize is tall and simple in morphology and structure,which is easy to describe and study.Taking maize colony as example,and with the aid of three-dimensional(3D)data acquisition equipment,we carry out the research of plant architecture parameter extraction,maize organ resource database construction,maize colony 3D modeling,mesh simplification,and canopy gap fraction calculation.These techniques are then applied to evaluate the architecture of maize plant and colony.The main work and innovations of this thesis are as follows:1.Phenotypic parameters extraction of maize based on 3D morphological data.Plant architecture parameters are mostly manual measured,which has the problems of low accuracy and inconsistent measurement standard of different persons.Image and 3D point cloud based parameter extraction could not meet the accuracy requirements of 3D maize modeling.By using the 3D digitizer acquiring 3D skeleton structure data and combined with the parameter definition,we extract the main maize plant architecture parameters,including plant height,leaf inclination angle,leaf azimuthal angle,and leaf length,etc.By solving the minimum plane angle of the difference sum function between each leaf azimuthal angle and the plane,we obtain the plant azimuthal plane.On the basis of which,a dev value describing the extent of leaf azimuth deviation is used for indicating plant compactness.The plant architecture parameters of maize extracted using 3D digitization could reflect the morphological differences between maize varieties and cultivation factors,and could be used in the 3D model construction of maize morphology.Point cloud segmentation of maize ear is conducted through calculating surface normal,shrinking transformation and Euclidean cluster.The segmentation rate is up to 90%,which provides key technical support for 3D maize ear analysis.2.Construction of 3D template resource database of maize organs.Currently,geometric modeling of maize plant and colony lacks highly accurate 3D organ templates.A method for constructing 3D template resource database of maize organ based on in-situ measured data is proposed.First,we draft 3D data acquisition standards of maize organs as the access rules of the database.Then,the 3D morphological data of maize organs are obtained according to the standard,and the geometric models of each organ are constructed based on the measured data,combined with 3D point cloud based reconstruction and parametric geometric modeling method of plant.Finally,the maize organ 3D template database is constructed by collecting the constructed geometric models of maize organs classified using visual and agronomic keywords.3.Maize colony 3D modeling method based on t-distribution.In order to build 3D maize colony models using a small amount of measured data,which could reflect the morphological and structural differences of caused by varieties,environmental conditions,management and cultivation measures,t-distribution functions of plant architecture parameters are constructed based on measured morphological parameters.And geometric models of maize organs in the resource database are called to match the generated plant parameters by t-distribution function by evaluating a similarity function.Then a geometric model of maize colony is constructed by combining the structure information of artificial interaction or image extraction.Finally,quadric error metrics based edge collapse algorithm and Laplacian smoothing operator based mesh simplification and optimization method is proposed,which provides a high quality mesh model for further visualization and analysis of maize colony.4.Canopy gap fraction calculation based on multi-scale subdivision hemisphere.Multi-scale hemispheres are constructed using(?)and Butterfly subdivision schemes.Most of the facets on the hemisphere are equilateral like triangles with very high similarity.Using these hemispheres and combined with Turtle model,canopy gap fraction is calculated with high accuracy and efficiency.The method is verified from the geometry and diffuse light distribution views respectively.The method provides key technical support for crop colony structure analysis,diffuse light distribution calculation and total amount of light interception simulation.5.Evaluation of maize plant architecture based on light distribution within the canopy.Combined with the crop canopy direct light and diffuse light distribution calculation methods,an instantaneous light distribution and total amount light interception calculation method is proposed by integrating the 3D colony modeling and canopy gap fraction calculation.On this basis,the plant architecture and light interception ability of several maize varieties are accurately and digitally evaluated.This paper constructed an analytical system for the morphological structure of maize colony in a digital way.It provides digitalization and visualization technology for the research and applications of productivity evaluation,high yields mechanism analysis and density tolerance of maize.