用于旋转对称三角传感器的一种鲁棒的处理方法(英文)

Robust processing method for rotationally symmetric triangulation sensor

旋转对称三角传感器由于其光学系统的旋转对称性而比其他的传统激光三角传感器具有更好的柔性,将旋转对称三角传感器应用于工业领域,相应的处理方法的性能是至关重要的。为了使传感器在复杂物体表面情况下仍具有较好的性能,我们设计了一种用于旋转对称三角传感器的鲁棒的处理方法。首先,分析旋转对称三角传感器图像的特性,这些特性取决于被测物体的表面特性,而这些图像特性将有助于传感器图像的处理方法的设计。其次,用一种由粗到精检测方法实现对传感器图像的处理。提出在精检测中,一种结合了自适应平滑和峰值法的方法,以实现检测的鲁棒性。在实际传感器图像上的实验结果证明,这种处理方法对于来自复杂表面的传感器图像具有很好的鲁棒性,从而提高了传感器的柔性。

The rotationally symmetric triangulation （RST） sensor has more flexibility than the traditional laser triangulation because of the rotational symmetrical optical system. For application of this RST sensor in the industrial field, the performance of processing method for measurement is very important. Aiming at the better performance on complex situation of object surface, a robust processing method is designed out here. Firstly, the sensor image property which depends on the characteristic of the object surface is analyzed for the design of the detection method.Then, the processing method is designed as a coarse-to-fine detection. In the processing method, a simplified Hough transform is used in the coarse detection to get a coarse radius; a new accurate and robust detection algorithm is prompted for the fine detection, which combines adaptive smoothing and zero-crossing of first-order derivative. The experiment result on the real sensor image shows that the processing method is robust to sensor image from the com- plex situation of object surface and has better performance on the real measurement.