星图中亮线的自动检测算法

Automatic Detection Algorithm of Bright Line in Star-Sky Image

对于星空观测CCD图像,当有高亮度恒星经过视场时,高亮度恒星成像使得CCD的读出电路电压过高,同一行(或列)像素灰度值偏高,图像中会形成过恒星中心的亮线.由于亮线和目标运动轨迹性质相似,对目标检测和识别的干扰较大,必须进行滤除.在对星图成像性质进行分析的基础上,提出了一种星图亮线的自动去除算法.利用恒星成像的频域各向同性性质和Fourier变换的自配准性质对亮线进行检测,并分别在空域和频域对亮线进行滤除.对三组观测星空序列图像(每组60帧)进行实验分析,结果表明均能成功地检测出星图中是否存在亮线.

For the CCD serial star images, if there is a high bright star in the field of view, the CCD readout circuit voltage can be over high and the same line or column gray value of the CCD image is on the high side, a light line traveling across the star centre is formed in the image. As the bright line exhibits great similarity to the motion of space targets, and disturbs greatly target detection and identification performance, it must be eliminated. Based on the analysis of the star-sky image, an automatic detection algorithm of bright line in star-sky images is put forward. With the isotropy property of star imaging and auto-registration character of Fourier trans-form, the algorithm can automatically detect and eliminate the bright line both in space domain and frequency domain. Experiments of 3 groups of star-sky images （60 frames per group） show that the algorithm can auto- matieally detect bright line with high accuracy.