Identification of crop weed based on image texture features

Jie Zhang¹, Fuli Song², Jiali Tang¹ 

¹School of Computer Engineering, Jiangsu University of Technology, Changzhou 213001, Jiangsu, China

²Personnel Division, Henan Radio & Television University, Zhengzhou 450008, Henan, China

By using computer image processing technology, texture features of weed in the corn seedling field are analysed, and then we present an algorithm combining Support Vector Machine (SVM) to form a classifier and promote an improved method of RBF network. The experimental results show that the proposed method is effective.