An adaptive machine vision system for full-field 3D surface profile measurement

3D surface profile measurement is important in manufacturing for process control and quality inspection. In this paper, a cost-effective machine vision system for adaptive and full-field 3D surface profiling applications is proposed. The system especially aims at the 3D inspection of small metallic objects that are the typical characteristics of electronic components such as weld solders and SMDs on PCBs. The proposed system is based on the phase shifting technique. A sinusoidal fringe pattern is projected on the object, and deformed in accordance with the object surface. The surface profile is obtained from the evaluation of intensity images of the deformed fringe patterns in different phase shifts. A software-controlled LCD panel is adopted to generate varying fringe patterns. It can shift the fringe pattern with accurate phase increments. During the phase shifting process it also eliminates vibration, which is the main error source in applying traditional phase shifting techniques. The system is adaptive because it can easily change the fringe period, phase-variation direction and projection area to accommodate the topography and reflectance of various objects. The measurement speed of the proposed system is fast. An image size of 640×480 pixels can be measured in less than 4 seconds with a typical personal computer. Experimental results from weld solders and SMT components on PCBs have shown the efficacy of the proposed method