基于机器视觉的现实叶片面积测量方法的研究

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。本文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，统计图像中目标象素点总数利用数学关系即可求出叶片面积。该处理方法大大降低了野外活体叶片面积的测量难度。

Research on Measurement of Realistic Planting Leaf Area Based on Machine Vision

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.