北京测绘2024，Vol.38Issue(12)：1754-1759,6.DOI:10.19580/j.cnki.1007-3000.2024.12.019
一种基于地基望远镜数据的目标检测方法
A target detection method based on ground-based telescope data
杨笑天 1李伟 1阴晓刚 1王梁 1谭金林 1吴意 1曹小敏1
作者信息
- 1. 西安空间无线电技术研究所,陕西 西安 710100||陕西航天技术应用研究院有限公司,陕西 西安 710100
- 折叠
摘要
Abstract
For ground-based telescope sequence images,this paper performed uneven background correction,fast image registration,and spatial satellite target detection.On this basis,the paper performed forward target association recognition on spatial satellite targets according to time series,extracted the motion trajectory of moving targets,and removed pseudo target points.A pruning strategy was proposed in the inter-frame association process of sequence images to delete erroneous trajectories,avoiding the problem of explosive dimensionality growth during the search process.The idea of reverse association was adopted for trajectory interruption in the case of missing frames.After the forward association was completed,the inter-frame target association recognition was carried out according to the reverse time series.Finally,the forward association recognition results and reverse association recognition results of each frame were merged to achieve association recognition of targets in ground-based telescope sequence images,reducing false alarm rates and improving detection rates.The results show that when there are many false alarm points,more frames indicate more accurate detection,and more targets contribute to obvious speed optimization.Through experimental verification,the detection results of this method are consistent with the known label results.关键词
背景不均匀校正/图像配准/图像去噪/目标检测/帧间关联Key words
uneven background correction/image registration/image denoising/target detection/inter-frame association分类
天文与地球科学引用本文复制引用
杨笑天,李伟,阴晓刚,王梁,谭金林,吴意,曹小敏..一种基于地基望远镜数据的目标检测方法[J].北京测绘,2024,38(12):1754-1759,6.
