Development of Image Analysis Based Methodology to Detect Weed in Drone Images

Weed infestation is a major problem faced in precision agriculture, and detecting infested area in a large field for its elimination is itself a very tedious task. Detecting weed patches via aerial images using digital image processing techniques is not much explored and relied-on method especially in the cases of overlapping color channels values of target and background and involvement of complex feature detection techniques, since normal feature detection methods do not provide clear outputs in such cases. In this paper, we have explored to develop an image processing-based methodology to detect weed with aerial image (UAV image) and we found quite encouraging results. This will also help in reducing the computational complexity.