基于星敏感器测试图像的噪声分析及仿真

Noise Analysis and Simulation Based on Test Images of Star Sensor

由于真实的星敏感器图像不易获取,仿真的星敏感器图像是测试的主要方式。仿真星敏感器图像对于噪声的添加大多采用高斯白噪声,缺乏对噪声水平的估算,与真实的星敏感器成像噪声有很大差别,使得星点定位,目标识别算法缺乏验证基础。为了模拟星敏感器的真实成像噪声,对星敏感器成像的噪声来源进行了全面分析,并以真实拍摄的星敏感器测试图像为基础,提供了相机参数的计算方法,同时对读出噪声和暗电流噪声的水平进行了测量,并给出了读出噪声、暗电流噪声、散粒噪声、热像素、宇宙线噪声和杂散光的仿真方法,利用实拍星敏感器测试图像的数据进行了仿真。结果表明,通过仿真的星敏感器图像噪声可以较好的还原星敏感器的真实成像,可为后续噪声滤波、星点定位、目标识别算法提供验证和测试。

Due to the actual star sensor image is not easy to obtain,using the simulated star sensor image is the primary way of testing.The simulated star sensor image mainly uses Gaussian white noise to add noise,lacking estimation of noise level,which is very different from the real star sensor image noise,making star point positioning and target recognition algorithm lack verification basis.In order to simulate the real imaging noise of the star sensor,this paper comprehensively analyzes the source of the star sensor imaging noise.It provides a calculation method of camera parameters based on the authentic star sensor test images.At the same time,the level of the readout noise and dark current noise is measured,and the simulation methods of readout noise,dark current noise,shot noise,hot pixels,cosmic ray noise,and stray light are given,and the simulation experiments are carried out by using the data of the authentic star sensor test images.The results show that the simulated star sensor image noise can better restore the actual image of the star sensor and can provide verification and testing for subsequent noise filtering,star point positioning,and target recognition algorithms.