CAJ | 학술논문

单目视觉定位在机器人视觉中具有重要的应用价值。结合光学成像原理、几何坐标变换以及相关的图像处理技术能够实现单目视觉目标定位。提出了一种基于RGB颜色空间的匹配滤波轮廓提取算法，进行特定目标的轮廓提取。该算法能精确、快速地计算出摄像机与目标物体之间的距离和角度信息，实现对特定目标的定位。基于NAO机器人平台的实际实验结果验证了该算法的有效性。
단목시각정위재궤기인시각중구유중요적응용개치。결합광학성상원리、궤하좌표변환이급상관적도상처리기술능구실현단목시각목표정위。제출료일충기우RGB안색공간적필배려파륜곽제취산법，진행특정목표적륜곽제취。해산법능정학、쾌속지계산출섭상궤여목표물체지간적거리화각도신식，실현대특정목표적정위。기우NAO궤기인평태적실제실험결과험증료해산법적유효성。
Monocular vision positioning has important application value in the robot vision. Monocular vision positioning can be achieved combining with the optical imaging principle, geometric coordinate transformation and image processing technology.In this paper,a particular target contour can be extracted by putting forward a kind of matched filtering con-tour extraction algorithm based on RGB color space. The distance and angle between the camera and the target object accurately and quickly can be calculated by the algorithm. Experimental results based on NAO robot demonstrate the ef-ficiency of the algorithm.