Improved phase detection algorithms for indirect time of flight range imaging cameras

Two significant sources of systematic error in current generation indirect time of flight range imaging cameras are systematic linearity error due to harmonics present in the system and axial motion error. Both of these errors are due to violations of assumptions made by the phase detection algorithm. This paper provides a theoretical framework to understand these errors and experimental results demonstrating that both of these errors can be mitigated by replacing the current state of the art phase detection algorithm with the Windowed Discrete Fourier Transform. Using this technique the RMS linearity error of the Victoria University Range Imaging System was decreased from 0.018 ± 0.002 m to 0.003 ± 0.001 m and the RMS motion error was decreased from 0.044 ± 0.002 m to 0.009 ± 0.004 m.