Dimensional error analysis in point cloud-based inspection using a non-contact method for data acquisition

The digital reconstruction of a computer aided design model from a physical object in the case of non-existence of such models has been a great challenge to engineers. With the advent of coordinate measuring machines (CMM), topological information of the points on the object's surface can be extracted in the form of a point cloud. However, CMM has been a very slow process with relatively low resolution which becomes a hurdle for a rapid manufacturing process. As a result, a non-contact method has been developed to provide higher speed and high resolution scanning. Therefore, the source of uncertainties for a non-contact probe, especially the laser scanner, had been reviewed to identify the potential improvement areas. Although the uncertainties are related to data acquisition capability, the surface construction effectiveness from the point clouds has similar importance to obtain the most accurate contour with minimum deviation from the actual surface. This can be achieved by eliminating the noise and by fitting the best shape according to the feature, such as a plane, cylinder or sphere, to the point cloud. The main aim of this work is to analyse and improve the performance of the laser scanner in terms of dimensional accuracy. A DIGIBOT II™ 3D laser digitizer which uses a remote ranging method was examined in the present work. The triangulation scanning system of this digitizer uses a high energy light source projected on the object's surface and a detector to sense the reflection. It was found that the scanning errors consist of random errors and systematic errors. For this scanner, the scanning data are random and follow the Gaussian distribution after the outliers are eliminated from the point cloud. The random error follows a systematic pattern where it varies with the scan depth and it is at a minimum when the y -coordinate is 228 mm and where the value of the error would be -0.0642 mm. The first systematic error is caused by the difference in the efficiency of photodetectors and the second systematic error is caused by the reflected laser intensity decrement as the scan depth increases. The second systematic error changes linearly when the y -coordinate is positive. Other methods for improving the scanning accuracy of the scanner have been recommended.