Weed Classification on Images using Machine Learning

Abstract

Classify the plant species images so that the machine distinguishes between crop seedlings and weed. The images include some stones in the background, we don't want the machine to learn the stones pixels and base its prediction on all the components of the seedlings + stones. Color-based segmentation: Keep just the green color pixels keep just the seedlings pixels. Clean plant images without stones in the background. Ready to be fed to the model. Weed control is considered as a major obstacle for the farmers in farming. The plant productivity decreases due to the poor control of the weed in farming. So inorder to increase the plant productivity we use the concept of weed classification.