轨迹测量系统中特征点检测技术研究

针对机器人直线轨迹测量系统中的图像处理鲁棒性问题,开展线结构光光条图像的特征点识别技术研究,提出了一种基于种子点拟合和直线生长的直线分割方法,很好地解决了各种条件下各段拟合直线的端点自动获取问题,提高了系统的环境适应性和测量精度.实验表明,该方法具有很好的鲁棒性.

In order to improve the robustness of image processing for robots' linear trajectory measurement system, the paper studies the technique of feature points' recognition in laser strip image based on line structured light, and presents a line segmentation method on the base of seed points fitting and line growing. It resolves the problem for automatically obtaining each fitting line end-points in various conditions, and improves adaptability to changing environment for robot work and precision. Many experiments show that the method has good robustness.