Enhancing Precision Agriculture: Computer Vision-Aided Farm Boundary Detection and Crop Land Identification From Drone-Captured Image

In the current technological era, Artificial Intelligence has played a significant role in addressing various challenges faced by people. Within the realm of Artificial Intelligence, computer vision emerges as a crucial subdomain with profound implications for agriculture. It is instrumental in tasks such as enhancing crop production, detecting diseases, and implementing precision and smart farming techniques. Unmanned Aerial Vehicles or drones represent another technological advancement that profoundly influences the collection of field data in agriculture. These devices navigate over designated areas, capturing images of fields, which are subsequently processed to extract pertinent information. This study leverages images acquired through drones, focusing on agricultural fields consisting of multiple plots. The research employs the Canny edge detection algorithm to segment the plots, and the Probabilistic Randomized Hough Transform algorithm to identify plot boundaries. Further, the work identifies areas within the plots where plants are present. To achieve this, the HSV (Hue, Saturation, Value) color space is utilized for processing drone images. The extracted boundary information holds paramount importance for autonomous UAVs and Unmanned Ground Vehicles. These vehicles utilize such information to optimize activities such as planting, harvesting, and pesticide or fertilizer application. Additionally, the delineation of plot boundaries offers farmers an accurate map of their fields, contributing to more efficient agricultural practices.