Study On Three Dimensional Reconstruction And Visualization Calculation Of Apple Tree Canopy

The apple tree is the largest fruit tree species in our country.The apple tree planting is an important part of agricultural production,and it is also an important source of the rural income.Appropriate tree structure,reasonable canopy light distribution play animportant role in improving the photosynthetic efficiency of fruit trees,promotingfruit yield,fruit quality and effective prevention and control of pests and diseases.Selection as well asreasonable improvementof tree crown distribution always is one of the core issues of pomology research,but the lack of quantitative calculation techniques and analysis methods as well as hardware and software tools related to tree canopy is still a majorproblem.In recent years,virtual plant has beenrapidlydeveloped as a multidisciplinaryly technology of botany,mathematics,computer graphics.Using 3D digital technology,the method of constrcting canopy three-dimensional structure model,conducting visualization computing and making precise analysis and evaluation based on measured data is an effective way to improve the scientific research and production of fruit trees in high quantitative level,and it is also one of the current hot and difficult fruit research issues,withimportant theory value and practice significance,at the same time,showing broad application prospects in the field of orchard automationmanagement,breeding,agricultural virtual training.In order to solve the calculation and analysis of 3D model of apple tree canopy based on quantitative calculation method,this dissertation mainly focuses on the apple tree canopy reconstruction,semantic feature representation,visualization computing,and integration of research and development of virtual pruning training software system,and the specific contents are as follows:(1)By means of 3D digitalizing,3D laser scanning technologies and equipment,combined with the manual measurement method,we have systemically acquired three-dimensional canopy structured data of an apple tree in a growth year,and established the multi-scale information integration based acquisition method of fruit tree morphological structure,to implement continuous observation and data acquisition of the anniversary growth of apple tree,providing accurate data base for 3D reconstruction of canopy.(2)We have presented a three dimensional reconstruction method of fruit tree canopy based on multi-scale measured data.First,the topological structure of dense point cloud was used to solve the problems of connectivity fault loop,splicing,deletion and small branches communicated with the exact solution branches radius in point cloud data of complex branch structure of the tree canopy,and to realize the accurate reconstruction of tree canopy deciduous period;then,according to the local and global features of crown point cloud,an elliptical hierarchical point cloud density shrinkage algorithm was implemented for separation of point cloud organ,so as to realize the rapid automatic reconstruction of canopy leaves;finally,a physics-based deformation and editing model of point cloud was constructed.The results show that the rate of correctly identified leaf number to the total leaf number of canopy is more than 90%,the rate of number of the leaves with angle deviation within 5 degree to the total leaf number of canopy is more than 95%.Finally,this paper presents a deformation algorithm of plant 3D scanning point cloud model based on physics,wind point cloud tree simulation results show that the deformation animation is smooth,and the method provides the technical support for the research of virtual wind simulation of fruit trees.(3)We have proposed a four layer domain graph-based semantic modeling and reconstruction method of tree canopy structure,including steps of analyzing hierarchical semantic features of different apple tree canopies,constructing semantic relationship model of layer,fractal,domain and directed tree of canopies by graph theory,constructing mathematical branches distribution the leaf distribution model based on the statistical theory,so as to realize the rapid reconstruction of three-dimensional structure of canopy.Comparing with the model of artificial measurement data,the 3D canopy models constructed by semantic model have the features that:the average error rate of total branch number was 8%,long branch and short branch ratio are less than 5%,the error of total leaf number less than 6%.The results show that themethod performs in a good adaptability,providing effective reference for tree pruning,analysis and experimental research virtual simulation.(4)Based on 3D canopy model of tree,we have constructed a Z-buffer algorithm-based direct sunlight and effective radiative transfer model and a Turtle algorithm-based sky scattering light and radiation transmission model,to implement light distribution calculation based on 3D tree canopy.We have validated the algorithm according to measured data,and the P value is 0.948 at a=0.05 level.We have conducted 3D visualization and analysis of light distribution in different canopy structures.By comparative analysis of canopy structures of three kinds of tree widely cultivated in China,the results show that the spatial distribution difference of canopies’ photosynthetic active radiation(PAR)receiving is obvious,and the occupancy is decreasing from top to bottom,as same as occupancy of leaf area.The distribution of PAR of opening tree structure concentrates upon central part because of its shape feature of center high and low crown.Contrarily,the distribution of PAR of evacuated layered and free fusiform form is uniform according to its branch distribution,and the rate of leaf area in upper part of canopy obviously increases.(5)Based on agent software architecture and technology,integrating modeling and light distribution computing method of apple tree canopy,we have developed the prototype software system of apple tree modeling and visualization computing(Apple_vcss),including modules of apple tree canopy reconstruction,semantic expression,visualization computing analysis,providing an interactive software tool for typical applications of virtual auxiliary pruning system.The results have shown that the proposed apple tree 3D modeling and 3D visualization computing algorithm could be effectively applied in the virtual pruning training system,providing real-time interactive process for the virtual pruning operations,visual display of pruning results and quantitative computing analysis of pruning decision,as a practical software tool for the research and the application field of orchard management,breeding,agricultural virtual training etc.