Weed Detection and Segmentation Using Deep Learning

In the recent times, where the automation systems has gained the highest priority to digitalize the world, the field of agriculture plays a major role in the growth of Indian economy. Weed plant detection and segmentation is a new research problem in the field of agriculture. In this paper we present a weed segmentation module. Weed classification module needs image processing task to be performed in order to detect the existence neural network to process the image and various classifiers such as random forest, Decision tree, SVM are used to classify the image the segmentation module makes use of U-Net architecture and Dense CRF is used for post processing in order to make boundaries of object more clear. The performance of the classifiers are measured using standard evaluation metrics.