Research on Improved YOLO11 Network with MSCA for Small Object Detection Education

The object detection algorithm based on YOLO series has been updated to YOLO11, which has a powerful image data learning ability and can complete a variety of functions based on images and videos. In order to better teach small object detection, in this paper we study the small object detection of oriented bounding boxes in YOLO11, combine an MSCA attention module, improve the network structure of YOLO11-OBB, and uses remote sensing image data set of Tianzhi cup competition for training and testing. The mAP performance of the improved algorithm framework is 1.7% higher than that of the original YOLO11-OBB algorithm on the N model, which realizes the optimization of the YOLO11 algorithm and can carry out small object detection more effectively. Finally, the improved algorithm is applied to the teaching activities of small object detection, and good teaching effects are achieved.