摘要

In fringe projection measurement system, the purpose of depth calibration is to establish a mapping relationship between fringe phase and object surface depth. In the calibration process, different data processing models and algorithms have significant influence on the measurement accuracy of the system. Focusing on this issue, this paper analyzes and compares the characteristics of several popularly used data processing methods in the calibration, including the error distribution characteristics of precision geometric model and its linearized model as well as polynomial approximation model in the presence of noise, and the impacts of node selection on the calibration precision of polynomial approximation model. Numerical simulation and experiment results verify the correctness of the analytic results.

全文