基于全景视觉机器人定位的路标提取方法

Landmark extraction method based on omni-vision robot positioning

针对全景图像像素点过多、图像复杂导致单一图像分割算法难以提取出图像中人工路标的问题,提出了一种用HSV阈值分割法与OTSU最大类间方差法相结合的算法.通过对两种算法的结合使用,可以更有效地滤除图像中的干扰区域及干扰点,从而将路标从图像中提取出来.利用三角定位法的相交圆法计算出移动机器人的坐标,完成对移动机器人的定位.结果表明,该方法能够提取出全景图像中的路标,有效地避免了错误提取的情况,具有一定的可行性.

In order to solve the problem that the single image segmentation method can hardly extract the artificial landmarks in the image due to the fact that the pixel points are abundant and the image is complicated in panoramic image, a method in combination with both HSV threshold segmentation method and OTSU method was proposed. Through the combination use of two methods, the disturbance areas and disturbance points in the image could be effectively eliminated, and the landmarks could be extracted from the image. In addition, the coordinate of mobile robot could be calculated with the intersecting circle method of triangulation positioning method, and the positioning of mobile robot was completed. The results show that the method can extract the landmarks in the panoramic image, effectively avoid the situation of error extraction, and show certain feasibility.