Font Size: a A A

Research On Potato Three-dimensional Phenotype Based On Computer Vision

Posted on:2021-12-22Degree:MasterType:Thesis
Country:ChinaCandidate:F F WeiFull Text:PDF
GTID:2493306011993589Subject:Master of Agriculture
Abstract/Summary:PDF Full Text Request
Digitalization and visualization of agriculture are the important research directions of agricultural researchers at home and abroad currently.With the continuous development of computer vision technology,people’s cognition of crops has deepened constantly in the 3D field.Crop models which are more realistic and 3D than images can be obtained by 3D reconstruction technology,sophisticated crop morphological data can be obstained by combining with 3D point cloud phenotype analysis technology,and its application in production practice will greatly promote the modernization of agriculture.Potato is the fourth most important food in the world.Although China’s potato output is high,the production technology is not mature,which seriously limits the development of potato production and its processed products.This paper analyzes and extracts potato morphological information by combining the image-based 3D reconstruction technology and point cloud processing technology,which has guiding significance for the intelligent development of potato industry.The main contents are as follows:(1)3D reconstruction of potato.An image sequence acquisition device based on image 3D reconstruction was built.A potato image sequence was collected using this device for 3D reconstruction,and a 3D point cloud model was established.(2)Research on multi-view 3D reconstruction.Connections between images,point clouds and 3D models were established.Batch 3D reconstruction of potatoes was processed by using Metashape software.(3)The morphological characteristics(length,width,height,volume,etc.)of potatoes were extracted based on the bounding box algorithm and Poisson reconstruction of potatoes point cloud shape information extraction method,and its effectiveness was analyzed by comparing with the measured data,the effectiveness of this method was verified.(4)Study different point cloud segmentation methods and apply them to potato point cloud segmentation.Combined with regional growth method and European clustering method,the curvature of potato point cloud is calculated and analyzed,the correlation between potato sprouts and curvature was verified.
Keywords/Search Tags:Computer Vision, Potato, 3D Reconstruction, Point Cloud Processing
PDF Full Text Request
Related items