Field Grass Navigation Dataset

The current agricultural industry faces challenges in crop health monitoring and weed management, particularly in large farmland areas where effective automated tools for detecting and classifying crops and weeds are lacking. Existing solutions often rely on manual monitoring, which is inefficient and prone to errors. Therefore, the Field Grass Navigation Dataset aims to provide high-quality annotated images to support research and applications in object detection and image segmentation. The dataset is constructed by capturing images in real farmland environments with high-resolution cameras, employing multiple rounds of annotation and expert reviews to ensure accuracy and consistency of annotations, and is stored in JPG format for convenient subsequent processing and analysis.