机械臂视觉伺服系统中的高精度实时特征提取

High accurate and real-time feature extraction in visual servoing systems for robotic arms

以机械臂视觉伺服系统为背景,立体视觉所需要的特征点为目标,提出一种基于颜色块的实时特征提取方法.特征点的定位分为粗定位和精定位两个过程,粗定位用来跟踪目标颜色块,精定位在粗定位的基础上,切出小图像并提取目标轮廓,利用K-L变换对轮廓点进行分类,并通过最小二乘法对边缘点进行拟合.根据直线方程计算顶点的亚像素精度坐标,为下一步的目标位姿测量提供基础.实验结果验证了该方法的有效性.

With the background of visual servoing for robotic arms and with the aim of features required by stereo vision, a method for real-time feature extraction based on the block of color is presented. The localization of feature includes coarse location and fine location. The coarse location is used to track the object. On the basis of coarse location, the fine location cuts a small image from the origin image, extracts the contour of the object from the small image and classifies the contour by K-L transform, then fits the edge points by using lease square error method. The sub-pixel coordinates of the feature points are calculated by the equations of lines, which provides foundation for the measurement of object＇s pose and position. The experimental results show the effectivenees of the method.