Phase Algorithms for Reducing Axial Motion and Linearity Error in Indirect Time of Flight Cameras

• Published in: IEEE sensors journal Vol. 13; no. 9; pp. 3386 - 3396
• Main Authors: Drayton, Benjamin M. M., Carnegie, Dale A., Dorrington, Adrian A.
• Format: Journal Article
• Language: English
• Published: IEEE 01.09.2013
• Subjects: algorithms; distance measurement; error correction; Robot vision systems
• ISSN: 1530-437X; 1558-1748
• DOI: 10.1109/JSEN.2013.2257737

Indirect time of flight cameras are increasingly being used in a variety of applications to provide real-time full field of view range measurements. Current generation cameras suffer from systematic linearity errors due to the influence of harmonics in the system and motion errors due to the requirement of taking multiple measurements. This paper demonstrates that replacing the standard phase detection algorithm with the windowed discrete Fourier transform can improve the root mean square (RMS) axial motion error with distance from 0.044±0.002 m to 0.009±0.004 m and the range from 0.112±0.007 m to 0.03±0.01 m for an object with a velocity of 2 m/s using a measurement time of 125 ms. This algorithm also improves the linearity of the camera by removing systematic errors due to harmonics, decreasing the RMS linearity error from 0.018±0.002 m to 0.003±0.001 m. This paper establishes the robustness of the windowed discrete Fourier transform, demonstrating that it effectively eliminates axial motion error over a variety of velocities and modulation frequencies. The potential for tailoring phase detection algorithms to specific applications is also demonstrated.