Weed recognition method based on color and morphological features in wheat field
综合颜色和形态特征的小麦田杂草识别方法

The technology of weed recognition based on machine vision becomes a hot issue of precision agriculture. Concerning the severe occluding of leaves of weed and wheat, a weed identification method was proposed with color and morphological features. Color feature was utilized to distinguish plants and background: a* was taken as characteristic variant in L*a*b* color space, and the improved method of maximum classes square error was taken as the criterion; Color feature was utilized to distinguish wheat and weed: hierarchical approach was used to color image segmentation in HSI color space; Morphological feature was utilized to obtain weed: using morphological opening and closing filter, and AND operation algorithm. The proposed methods together with a chemical weeding system were simulated and the efficiency of the overall system was evaluated theoretically. The experimental results on a series of weed images show that the correct identification rate exceeds 92.6%, and the herbicide reduction rate exceeds 72.4%.