关键词:
in-orbit target detection
lightweight
remote sensing
RSOD dataset
YOLOX
ZCU102
摘要:
With the rapid development of space remote sensing technology, the number of high-resolution optical remote sensing images is increasing exponentially. The detection of high-strategic-value targets, such as aircraft, is currently a hot research topic in the field of high-resolution image processing for remote sensing. Traditional remote sensing image object detection algorithms include template matching and traditional machine learning. These algorithms mostly rely on prior knowledge from experts, and the features used are generally primary manual features limited to the pixel level, so they have certain limitations and cannot cope with complex, ever-changing backgrounds and diverse multimodal targets. Remote sensing object detection algorithms for deep learning technology include two- and one-stage methods. Two-stage methods have high accuracy, but they consume abundant resources and have limited processing speeds. YOLO detection algorithms elicit much concern and are applied because of their simple network structure and balanced detection accuracy and speed. However, one-stage models cannot be directly deployed on embedded devices in satellites for real-time detection of aircraft targets because of the limitations in computing power, storage capacity, and model complexity. Therefore, lightweight network models that have reduced demands for computing power and storage need to be developed. These network models with excellent target detection capabilities can then be deployed to aerospace chips with limited resources to complete efficient aircraft target detection tasks. To address the difficulty of deploying current network models with excellent target detection capabilities to aerospace chips with limited computing and storage resources, this study proposes five targeted designs for the benchmark model, which is based on the one-stage YOLOX-s algorithm, and the implementation of the model adopts a lightweight design concept. A fast optical remote sensing aircraft