论文部分内容阅读
精确的成像模型是星敏感器高精度姿态测量的关键,尤其是成像模型中镜头畸变模型的选择。根据几何推断的最大信息量准则,对不同的镜头畸变模型在不同星点位置的噪声水平以及不同畸变系数情况进行了仿真比较,并通过实验进行了验证。实验结果与仿真分析结果一致,结果表明:在镜头畸变系数小于2e-5且星点位置噪声水平小于0.5像素时,采用一阶径向畸变模型即可;当镜头畸变系数大于2e-5或星点位置噪声水平大于0.5像素时,可采用3参数的一阶径向畸变加一阶切向畸变,更多参数及更高阶模型的优势并不明显。
The accurate imaging model is the key to high-precision attitude measurement of star sensor, especially the choice of lens distortion model in imaging model. According to the maximum information criterion deduced by geometry, the noise level and different distortion coefficients of different lens distortion models at different positions are simulated and compared, and verified by experiments. The experimental results are consistent with the simulation results. The results show that the first order radial distortion model can be used when the lens distortion coefficient is less than 2e-5 and the noise level at the star point is less than 0.5 pixels. When the lens distortion coefficient is greater than 2e-5 or The first-order radial distortion plus the first-order tangential distortion of the three parameters can be adopted when the noise level at the point location is greater than 0.5 pixel. The advantages of more parameters and higher-order models are not obvious.