基于色标和尺度不变特征的实时特征提取方法

摘要

With the background of mobile robot vision and with the aim of features required by monocular vision, a method for realtime feature extraction on the base of block of color and the Scale Invariant Feature Transform (SIFT) feature point operator is presented.The localization of feature includes color labeling location and feature points location. The color labeling location is aim to find the centre of gravity point of the color label from the scaling picture. The feature points location is on the base of color labeling {ocation, cuts a small image from the origin image, extracts the SIFT feature points of the object from the small image. The position of the tracked object are calculated by the max or min feature points, it provides foundation for the object tracking. The experimental results show it's effective.%以移动机器人视觉系统为背景,以单目视觉所需要的特征点为目标,提出一种基于颜色块和尺度不变特征点算子的实时特征提取方法;目标的定位分为色标定位和特征点定位两个过程,色标定位用来寻找在缩变图像上目标颜色块的重心点,特征点定位是在色标定位的基础上,切出小图像并提取目标的尺度不变特征点,根据极值特征点计算目标位置,为下一步的目标跟踪提供基础;实验结果验证了方法的有效性.