Measurement of Plant Leaf Area Based on Computer Vision

Abstract

It is laboring and tedious to measure plant's leaf area. For this problem, a novel approach based on computer vision was proposed for measuring leaf area. Firstly, plant leaves image were captured by image scanner and transformed to HSI color space. Secondly, combining with saturation, leaves were segmented from background with Otsu approach based on Hue information. Then, morphological filtering was performed to eliminate most noises. After that, blob analysis was used to fill the holes in the leaves and extract the leaves to measure area from original image. Experiments showed that the proposed approach had better segmentation result than the indices-based ones. Moreover, the proposed approach was proofed to be fast and efficient to measuring leaf area comparing with leaf area meter.