论文部分内容阅读
为了保证工业机器人磨抛的加工质量,利用激光扫描技术对机器人夹持工件的形状误差以及装夹误差进行测量和评估,包括点云数据的获取和去噪。采用条纹式激光扫描仪配合直线匀速运动对机器人末端夹持工件进行扫描,通过调节测量和运动参数,获取近似网格点云。为了去除点云中存在的大尺度噪点,在K近邻均值滤波(KNNMF)算法基础上,提出了基于局部均值的K近邻均值滤波(LMKMF)算法对偏大的数据点进行局部预先滤波,并建立相关数学模型。以峰值信噪比作为评价标准,以实际测量点云样本为测试对象进行去噪测试。结果表明,相比标准的KNNMF算法,结合LMKMF预先滤波的KNNMF算法在30%噪点密度下去噪能力提升了53.78%,证实了其在高密度噪点下具有更好的去噪能力和特征保持能力。
In order to ensure the machining quality of industrial robot grinding and polishing, laser scanning technology is used to measure and evaluate the shape error and clamping error of the robot clamping workpiece, including the acquisition and denoising of point cloud data. A stripe laser scanner was used to scan the end of the robot for clamping the work piece along with a linear uniform motion. The approximate grid point cloud was acquired by adjusting the measurement and movement parameters. In order to remove the large-scale noise existing in point clouds, a K-nearest neighbor mean filter (LMKMF) algorithm based on local mean is proposed based on K-nearest neighbor mean filter (KNNMF) algorithm, which locally pre-filters oversized data points and establishes Related mathematical models. Peak signal to noise ratio as the evaluation criteria, the actual measurement point cloud samples for the test object to test noise. Compared with the standard KNNMF algorithm, the denoising capability of KNNMF with LMKMF pre-filtering is improved by 53.78% at 30% noise density compared with the standard KNNMF algorithm, which proves that it has better denoising ability and feature preserving ability under high density noise.