测量机器人小视场星图一维最大熵星点图像分割算法

摘要

As one of the fundamental problems in processing star map,image segmentation plays a significant part in ensuring precise field astronomical survey.Image binarization is the key procedure in the image segmentation,but it is extremely difficult to extract star targets from complex sky background using conventional threshold segmentation algorithms.Considering that the Leica video measurement robot TS50i shows features such as the small field of view,single star point,weak target,and single peak,one-dimensional maximum entropy method is firstly proposed to split the star maps.The proposed algorithm is verified by comparison with conventional threshold segmentation algorithms.It is indicated that the one-dimensional maximum entropy algorithm can achieve satisfied binarization processing results while adequately preserve the image information at the same time.Simulation experiments using real star maps show that the extraction method based on this algorithm is accurate and reliable with an accuracy of an order of magnitude better than requirements of the field first-class astronomical survey,hence it can satisfy the need of precise field astronomical survey.%图像分割是处理恒星星图的基本问题之一,也是确保野外高精度天文测量精度的关键步骤.图像分割的重点是图像二值化,对于星图的复杂星空背景,传统的阈值分割算法难以将星点目标从背景中提取出来.针对徕卡视频测量机器人TS50i拍摄恒星星图的小视场、单星点、弱目标和单峰性的特点,首次提出了运用一维最大熵法对星图进行阈值分割.对比分析了常用的几种阈值分割算法,通过对分割结果的比较和定量分析,验证了一维最大熵法能够在充分保留图像信息的同时,对TS50i图像达到具有良好的二值化处理效果.基于大量实拍星图背景的仿真试验表明,以该算法为基础的星点提取算法的准确性和可靠性得到验证,星点提取精度较野外一等天文测量精度要求高一个数量级,可以满足野外高精度天文测量的需求.