关键词: 空间小目标, 复杂场景, 成像模型, 图像处理, 动态仿真

Abstract:

In response to the universal demand for space target detection technology research in space image data sources, this study focuses on the problems of insufficient training data for intelligent algorithms and the use of single data for traditional algorithms, with the goal of generating dynamic digital sequence images of small space targets in complex scenes. A visible light digital imaging simulation system based on a space observation platform is designed. A small target imaging model is proposed, which is based on two-dimensional shape feature point description and imaging analysis modelto carry out digital modeling and imaging simulation of space targets. A background star imaging method is designed. Through the selection of stars in the field of view, the transformation of imaging coordinates and the sub-pixel position modeling, the accuracy of the star point target position imaging modeling can reach ±0.005 pixels. The dynamic models of various light backgrounds in space environmentare designed, and the imaging noise analysis and modeling are carried out. The types of scene and the test conditions of robustness detection of space target images are topologies. The imaging simulation functions such as dynamic change of target graylevel, target curve movement, motion trailing and platform jitter are added to the design of motion simulation algorithm. The simulation results show that the simulation system can generate customized dynamic digital image sequences. After the necessity and adequacy verification and evaluation, the digital simulation system can meet the demand of multiple image data for the development of spatial object detection algorithm, and has certain practical value for the research of spatial object detection algorithm.

Key words: space small target, complex scene, imaging model, image processing, dynamic simulation