Abstract:

Much noise exists in cloud data due to deviation of system and matching algorithm in photogrammetry. This paper presented an algorithm for scan line cloud data filtering. It fulfiled the noise filtering, and rectified data in creep region. Experiments show that it has higher speed and better results than that of comparison algorithm. The filtered results are most applicable to curve or surface reconstruction.

Key words: photogrammetry, scan line cloud data, morphological operator, open operation, close operation, adaptive noise filtering

摘要：

摄影测量的点云数据通常叠加有多种形式的噪声信号，针对摄影测量中得到的扫描线形式的点云数据，采用形态学算子进行了噪声滤除。在完成噪声滤波的同时，在一定程度上对零件陡峭处获得的数据进行自动拟合。实例验证表明，该滤波方法处理速度快，其效果比其他的滤波方法好，滤波后的数据符合曲线曲面重构的要求。

关键词: 摄影测量, 扫描线点云数据, 形态学算子, 开运算, 闭运算, 自适应噪声滤波