Speckle noise, dynamic range of light intensity, and spurious reflections are major challenges when laser scanners are used for\n3D surface acquisition. In this work, a series of image processing operations, that is, Spatial Compound Imaging, High Dynamic\nRange Extension, Gray Level Transformation, and Most Similar Nearest Neighbor are proposed to overcome the challenges coming\nfrom the target surface. A prototype scanner for metallic surfaces is designed to explore combinations of these image processing\noperations.Themain goal is to find the combination of operations thatwill lead to the highest possible robustness and measurement\nprecision at the lowest possible computational load. Inspection of metallic tools where the surface of its edge must be measured at\nmicrometer precision is our test case. Precision of heightsmeasured without using the proposed image processing is firstly analyzed\nto be �±7.6
Loading....