基于双目立体视觉监测玉米叶片偏转运动方法研究

项目名称： 基于双目立体视觉监测玉米叶片偏转运动方法研究

项目编号： No.31501226

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 农业科学

项目作者： 王传宇

作者单位： 北京市农林科学院

项目金额： 20万元

中文摘要： 玉米叶片偏转运动的定量化监测对于评价高产群体冠层结构有重要意义。具有高时空分辨率的连续在线监测方法是国内外共同关注的热点，本项目选择双目立体视觉技术作为叶片偏转运动定量化获取方法，针对双目立体视觉技术在农作物生长监测上易受环境和作物本身影响等问题，重点研究变化光照条件下玉米叶片图像分割方法，复杂情景下玉米叶片的立体匹配与三维重建方法，建立叶片偏转角度与冠层光环境的定量化模型。为评价高产品种的耐密性提供方法途径，为超高产栽培技术体系提供理论支撑，促进农学、信息学、机电工程学之间的学科交叉和融合。

中文关键词： 作物长势；形态信息；图像采集；双目立体视觉；三维重建

英文摘要： Quantitative monitoring of the azimuthal movement of maize blade plays an important role in the evaluation of high yield canopy structure. Continuous on-line monitoring methods with high spatial and temporal resolution has been become a hot topic around the world. The project selects binocular stereo vision technology as azimuthal movement of maize blade quantitative acquirement method, as to the problem of binocular stereo vision technology on crops growth monitoring is vulnerable to environment and crop itself, the major concern of project is maize leaf image segmentation under illumination invariant conditions, stereo matching and 3D Reconstruction of maize leaves under complex context, establishing a quantitative model of blade azimuthal movement angle and canopy light environment. The project provides an approach to evaluate the Density-tolerance of high-yielding varieties ,provides theoretical support for ultra-high yield cultivation technology system, promotes integration of agronomy, information science, electrical and mechanical engineering.

英文关键词： Crop growth condition;Morphological information;Image acquisition;Binocular stereo vision;3D reconstruction