相机检校中标志点的自动提取  

Auto-extraction of mark points on camera calibration

作　　者：李国胜[1,2,3] 林宗坚[3] 任超锋[3,4] 彭晓东[3,4] 

机构地区：[1]中国矿业大学环境与测绘学院,江苏徐州221116 [2]中国石油大学地球资源与信息学院,山东青岛266555 [3]中国测绘科学研究院,北京100039 [4]武汉大学遥感信息工程学院,武汉430079

出　　处：《测绘科学》2010年第6期161-163,共3页Science of Surveying and Mapping

基　　金：“十一五”国家科技支撑计划项目(2006BAB15B01);863计划重点项目(2008AA121305);中国气象局气象新技术推广项目(CMATG2010Z06)

摘　　要：本文针对相机检校中,手工选择标志点速度慢、强度大的缺点,提出了基于边缘提取和形状系数多级约束的方法。首先对图像进行Wallis滤波,并对滤波后的图像采用Canny算子进行边缘提取;然后在八邻域跟踪的基础上,根据边缘长度、边缘首末端点距离、形状系数和光谱信息检测符合条件的标志点边缘。通过实验验证,该方法可提取出92%左右的标志点。Traditional method manually selected marking sign is slow speed and laborious on camera calibration,so a way was proposed based on edge extraction and shape coefficient in the paper.Firstly,Wallis filter was applied to image and edge was extracted based on Canny algorithm to Wallis filtered image.Then,through eight neighborhood track,edge of marking sign was detected based on edge length,distance of x-axis,y-axis of head and end node,shape coefficient and spectral information.The experiment was carried out,92% marking signs were extracted by this way.

关 键 词：标志点 八邻域跟踪 WALLIS CANNY 形状系数 

分 类 号：P246.1[天文地球—测绘科学与技术]