【摘要】 在计算机视觉测量中,摄像机的标定精度对保证系统测量精度有着至关重要的作用。而角点的提取精度一直是摄像机标定技术的难点。文中在分析常用角点提取方法的基础上,结合数学形态学算法和边缘形态模板,施行了畸形角点的修正,实现了高位置精度的角点提取,并在分析算法和计算结果的基础上,提出了角点位置突变概念,进行了亚像素级的角点像素坐标修正。标定中的角点经模板校正、数学形态学提取、亚像素修正可达到很高的精度。该算法计算简单、灵活,程序自动化,简化了标定过程。更多还原

【Abstract】 The accuracy of camera′s calibration is crucial to the reliability of measurement in the computer vision system.The precision of corner detection is always a difficult problem.Based on the analysis of some common algorithms for corner detection,and combined with mathematical morphology and the edge appearance template,this paper practiced to the correction of the abnormal corner and carried out corner detection of high position accuracy through experiment.With the foundation in analytical algorithms and calculation result,this paper put forward the position mutation concept and the sub pixel modified method.After the edge appearance template correct,corner detection based on mathematical morphology,and sub pixel correction,the corner detection can attain very high accuracy.This algorithm is simple,vivid,procedure automation and it can simplify the calibration process.更多还原