Abstract：The target information from the high-resolution remote sensing images using unmanned aerial vehicle(UAV) platforms is critical to military action planning. However, the large size and mixed background information of high-resolution remote sensing images can cause a lot of false alarms and misidentifications in the target information detection results. In addition, due to the huge difference between the image and the target size, detecting small targets in a large number of background pixels must employ a higher accuracy detection algorithm. To solve the above problems, this paper proposed a tiny target detection algorithm utilizing the block information of UAV remote sensing images. The model first alleviated the processing difficulties and slow speed caused by large image sizes using block information and a small neural network structure. Then, the global attention mechanism was applied to suppress possible false alarms. Finally, the detector and the classifier information during the training process was exchanged to improve the capabilities of both at the same time. The proposed detection algorithm was verified on a large-size remote sensing image dataset. The results show that the number of false positives in the detection results is significantly reduced and the accuracy is greatly improved, demonstrating the effectiveness of the proposed method.
李楚晨, 唐善军, 赵冰青. 一种基于无人机探测图像区块信息的弱小目标检测算法[J]. 空天防御, 2025, 8(1): 41-47.
LI Chuchen, TANG Shanjun, ZHAO Bingqing. Weak Object Detection Algorithm Based on High Resolution Remote Sensing Image of UAV Platform. Air & Space Defense, 2025, 8(1): 41-47.
