Automatic Segmentation Algorithm for Wheat Field Images Based on UAV

High-resolution Unmanned Aerial Vehicle (UAV) imagery plays a pivotal role in agricultural management and surveillance. The precise and efficient processing of these images is paramount. This research introduces an automated image segmentation algorithm designed for the segmentation of UAV-captured images from wheat test fields. Initially, the UAV-acquired images undergo rectification through Radon transform to enhance image quality. Subsequently, the Region of Interest (ROI) within the UAV image is extracted using the Extra Green algorithm (EXG) to eliminate superfluous elements and reduce computational complexity. Finally, the target image is obtained and labeled in accordance with specified criteria using multiple segmentation techniques. This study explores three UAV flight altitudes (15, 45, and 91 m) for capturing images of the wheat test field. The acquired data at varying altitudes undergo automatic segmentation. The outcomes demonstrate that the automatic segmentation algorithm proposed in this study exhibits a notable degree of accuracy in segmenting wheat field images at different altitudes. This advancement furnishes a potent tool for facilitating intelligent decision-making and resource allocation in the realm of agriculture.