A Survey on Image Processing Methodologies for Crop and Weed Detection

Abstract

Artificial intelligence is nowadays a fast-growing area of study, in particular deep learning. One of its different uses is object recognition and computer vision. This work is intended by the integration of the two technologies. This work created a system to identify various plants and weeds as an alternative to the system on FarmBot robots. This is achieved via access to the images through the FarmBot API, computer vision for the processing of the image, and artificial intelligence for transmission learning through an RCNN that autonomously carries out plants. In addition, weeds coordinates are given as results. The machine output is compared to related research study as well as the latest version of the FarmBot weed-detector. This study forms a technical viewpoint and offers an alternative to conventional agricultural weed detectors, opening doors to intelligent and sophisticated systems.