A Method Of 3D Vector Modeling Based On Computer Vision And Analyzing For Plant Root

Computer vision,also known as machine vision,refers to the use of cameras and computers to simulate human vision to identify,track,measure,etc.machine vision,and further processing through recognition and analysis to make it more suitable The human eye observes or transmits to the image detected by the instrument.Computer vision technology should not only be an extension of the human eye,but also should be part of the function of the human brain.For example,the three-dimensional model of the acquired two-dimensional image is constructed,and the artificial intelligence technology is used to complete some difficult intelligent applications or specific functions.Computer vision technology is a key technology that cannot be obtained.Currently,computer vision Technology has replaced many manual tasks,such as license plate recognition,face recognition,automatic driving,etc.,but in agriculture,remote sensing technology,crop dynamic monitoring,disease prevention,agricultural product quality grading and other fields have a wide range of applications.At present,the research on the three-dimensional configuration of plant roots is roughly divided into two types,one is complete in situ information,that is,to ensure the state of plants in the soil,and then directly obtain the plants by using magnetic resonance imaging(MRI).Three-dimensional configuration,but this method has great influence on the three-dimensional structure of roots due to intricate soil internal information and a large amount of soil medium,and can not accurately reflect the three-dimensional configuration of roots,but this method ensures the integrity of plants.Sex.The second is to separate the root system from the soil.After simple pretreatment,the obtained root structure is placed in the laboratory environment,and the three-dimensional model of the root system is obtained by using a three-dimensional scanning device such as a CT scanning device.Currently,the method is also the mainstream to obtain the root system.The method of three-dimensional configuration,but this method has high requirements on laboratory equipment,and the equipment is expensive,and can not be easily and efficiently extended to various fields,and this paper proposes a new method that uses computer vision technology to separate After the root system is cleaned,in the experiment,the camera is used to photograph the different angles of the target object,and then the computer technology and the corresponding algorithm are used to obtain the 3D model of the plant conveniently and concisely,and then complete on the basis of the visual 3D model.Measurement of individual phenotypic data for roots.The research content of this paper is as follows:(1)A detailed understanding of the domestic and international research on computer vision and the three-dimensional configuration of the root system,analysis and comparison of the advantages and disadvantages of each process algorithm,and determine the technical route of the research work.(2)Using the camera to complete the acquisition of the three-dimensional data of the target root,and then first performing a series of preprocessing,such as filtering,and enhancement processing on the acquired two-dimensional sequence image,reducing the influencing factors on the single image,and ensuring the later three-dimensional The accuracy of the modeling,then camera calibration of the camera system,find the internal parameters and external parameters,provide the necessary data for the later three-dimensional modeling,and finally select the optimal algorithm to complete the construction of the root model.(3)Optimize the construction algorithm.In the process of 3D construction,due to the need to voxelize the bounding box,if the selected 3D resolution is too high,although the details of the root system can be completely reflected,but the construction is completed.The time spent is exponentially increasing.Compared to some common algorithm ideas,such as the octree algorithm,the construction process is optimized to save computation time.(4)System implementation,using the visualization platform to display the acquired 3D data,and acquiring the phenotypic data of the root 3D model by means of mouse picking,and comparing the measurement results.