高分辨率遥感影像的圆特征提取广泛应用于摄影测量中,其精度高低对相机标定和特殊目标监测有着重大的影响。在实际应用中,圆特征提取的精度高低主要取决于所使用的圆特征提取算法。本文研究总结了重心法、中心收缩法和基于边缘跟踪的最小二乘圆拟合算法等常用的圆特征提取算法的原理和实现步骤,并根据实验数据的比较分析,给出了这3种常用圆特征提取算法各自的优缺点和适用情况。最后,提出并实现了一种基于边缘二次跟踪的双圆检测算法,解决了由成像目标缺损引起的提取圆心位置偏差的问题。
Circular feature extraction of the high resolution image has been widely used in photogrammetry,and the precision of the extraction has a great effect on digital camera calibration and specific target monitoring. Based on the summarize of 3 common circular feature extraction algorithms,according to the results of the tests,this paper calculated that the advantages and disadvantages of the 3 different circular feature extraction algorithms,include gravity algorithm、center shrink algorithm and least squares analysis algorithm based on edge tracking. At last,this paper suggested a double circle extraction algorithm,solved the problem of inaccurate center position caused by deviation of the target.