Rapid Automatic Multiple Moving Objects Detection Method Based on Feature Extraction from Images with Non-sidereal Tracking
Lei Wang,Xiaoming Zhang,Chunhai Bai,Haiwen Xie,Juan Li,Jiayi Ge,Jianfeng Wang,Xianqun Zeng,Jiantao Sun,Xiaojun Jiang
2024-09-04
Abstract:Optically observing and monitoring moving objects, both natural and artificial, is important to human space security. Non-sidereal tracking can improve the system's limiting magnitude for moving objects, which benefits the surveillance. However, images with non-sidereal tracking include complex background, as well as objects with different brightness and moving mode, posing a significant challenge for accurate multi-object detection in such images, especially in wide field of view (WFOV) telescope images. To achieve a higher detection precision in a higher speed, we proposed a novel object detection method, which combines the source feature extraction and the neural network. First, our method extracts object features from optical images such as centroid, shape, and flux. Then it conducts a naive labeling based on those features to distinguish moving objects from stars. After balancing the labeled data, we employ it to train a neural network aimed at creating a classification model for point-like and streak-like objects. Ultimately, based on the neural network model's classification outcomes, moving objects whose motion modes consistent with the tracked objects are detected via track association, while objects with different motion modes are detected using morphological statistics. The validation, based on the space objects images captured in target tracking mode with the 1-meter telescope at Nanshan, Xinjiang Astronomical Observatory, demonstrates that our method achieves 94.72% detection accuracy with merely 5.02% false alarm rate, and a processing time of 0.66s per frame. Consequently, our method can rapidly and accurately detect objects with different motion modes from wide-field images with non-sidereal tracking.
Instrumentation and Methods for Astrophysics