LEBR-YOLO: Accurate Fruit and Vegetable Detection Method for Space Station Cargo Bay

In order to address the issue of detection accuracy, which is compromised by hand occlusion when astronauts retrieve items from the space station cargo bay, a modified algorithm has been developed. This algorithm, termed “YOLO11,” is designed to enhance the detection of fruits and vegetables in the cargo bay environment. LEBR-YOLO. The method enhances the initial convolution to a lightweight structure, thereby reducing the computational effort while expanding the receptive field. Additionally, a two-layer attention mechanism is incorporated to enhance the ability to extract features. The introduction of an enhanced lightweight shared deformable detection module has been demonstrated to enhance the detection capability in the presence of occlusion. Experimental results demonstrate that the proposed method attains 94.3 % accuracy, 88.5 % recall, and 93.3% mAP50 on the proprietary fruit and vegetable dataset, while preserving low model complexity to satisfy the requirements of the Lightship cargo spacecraft in orbit. The proposed method has been shown to be effective in the detection of fruit and vegetable items in the space station, enhancing detection accuracy and reducing false detections and omissions.