一种基于Sobel算子的星敏感器星图预处理方法

星敏感器在太空环境下工作时容易受到杂散光的干扰,对星图进行二值化后直接采用质心法难以从星图中提取恒星星像坐标.针对上述问题,提出了一种基于Sobel算子的星敏感器星图预处理方法,用于提取星像坐标.首先采用Sobel算子计算星图中所有像素的梯度,保留所有大于阈值的像素梯度作为新的星图,然后以新的星图作为样板通过质心法从原始星图中提取恒星星像.由于杂散光干扰信号在星图中表现为缓慢信号,这些信号计算的梯度很小,因此通过Sobel算子计算后新的星图只有恒星星像.该方法能够消除星图中的杂散光干扰信号,使星图中只保留恒星星像信息,提高恒星星像坐标提取精度,具有较好的鲁棒性.

A Star Image Pre-processing Method of Star Sensor Based on Sobel Operator

Star sensors are vulnerable to the interference of stray light when working in space. It is difficult to extract the coordinates of the fixed star image from the star image by means of the centroid method after the binaryzation of star image. To solve the above problem, a novel star image pre-processing method for extracting start image coordinates of star sensor based on Sobel operator is presented in this paper. Firstly, Sobel operator is used to calculate the gradient of all pixels in the star image. All pixel gradients larger than the threshold are retained as a new star image and then the new star image is used as the template to extract the fixed star image from the original star image by the centroid method. Because the stray light interference signals behave as slow signals in the star image, the gradient calculated by these signals is very small, the new star image calculated with Sobel operator is only the fixed star image. The proposed method has good robustness because it can eliminate the stray light interference signals in star image and improve the accuracy of extracting fixed star image coordinates.