Detecting Crops and Weeds in Fields Using YOLOv6 and Faster R-CNN Object Detection Models

摘要

It is true that every nation's progress, particularly India's, depends heavily on its agricultural sector. Farmers are always concerned about weeds because they may dramatically lower agricultural output and cause financial losses. Therefore, the productivity and profitability of the agricultural industry may be significantly impacted by the development of technologies that can precisely detect and control weeds. Robotic weed detection and eradication is a promising technology that can reduce labour costs and time spent while increasing weed control's precision and efficacy. Robots must, however, be outfitted with sophisticated computer vision systems that can precisely identify and categorize items in real-time if they are to properly discriminate between crops and weeds. Two well-liked methods for object recognition in computer vision are Faster Region - Convolutional Neural Network (Faster R-CNN) and You Only Look Once (YOLO). Our study can assist in determining the best algorithm for weed detection and management by evaluating their performance on a sizable collection of photos of crops and weeds. Accurate weed identification and management can also be advantageous for precision agriculture, which makes use of technology and data analytics to improve agricultural output. Farmers may use less pesticides, saving money and reducing their impact on the environment, by more accurately focusing on weed-infested regions. Overall, our study has the potential to develop agricultural technology, increase crop yields, and guarantee food security for the world's expanding population.