Accurate three-dimensional measurement for small objects based on the thin-lens model

Both accuracy and depth of field (DOF) are required for small objects' measurement in advanced manufacture and accurate robotics industries. In this paper, a stereo vision system with structured light based on the thin-lens model is developed to measure small objects with high accuracy and extended DOF at high magnification. The model of the proposed stereo vision system is built. The new system calibration scheme and measurement procedures are proposed. The DOF of the used thin-lens model is extended by utilizing autofocus capability. With the thin-lens model, accurate calibration and extended DOF at high magnification can be achieved. Three measurement experiments are conducted with the proposed system to evaluate its effectiveness and accuracy. The error of the protrusions' height on the reconstructed aluminum part is lower than 0.55%, and the standard deviation (STD) of a fitting plane reaches 1.7 µm. Low cost, high accuracy, and extended DOF can be simultaneously achieved for measuring small objects with the proposed system based on the thin-lens model.