基于机器视觉的现实叶片面积测量方法的研究  

Research on measurement of realistic planting leaf area based on machine vision

作　　者：陈鼎才[1] 王定成[2] 查金水[1] 

机构地区：[1]中国科学院合肥智能机械研究所,安徽合肥230031 [2]南京信息工程大学计算机科学与技术系,江苏南京210044

出　　处：《计算机应用》2006年第5期1226-1228,共3页journal of Computer Applications

基　　金：国家863计划资助项目(2003AA118070)

摘　　要：叶面积是一些作物栽培中常用的指标,也是农作物的产量和品质的评价指标,而现实中获取的叶片图像都具有一定程度的几何畸变。文中研究了利用机器视觉技术对有透视畸变的叶片图像进行几何校正,使校正后图像尽量接近于叶片在无透视畸变情况下能获得的图像,统计图像中目标像素点总数利用数学关系即可求出叶片面积。Leaf area is a common criterion in some crop plantings, and it is also an evaluating criterion for quantities and qualities of crop products. But leaf images got from the real world often have some geometric distortion. Now the geometry correction to leaf images with perspective distortion was researched by knowledge of machine vision, and the objective is to make the corrected leaf images be close to the leaf images without geometric distortion as far as possible. Then the leaf area can be calculated with the math relations according to the statistic of goal pixel.

关 键 词：机器视觉 透视畸变 几何校正 叶片面积 

分 类 号：TP242.62[自动化与计算机技术—检测技术与自动化装置]