An image-based calibration technique of spatial domain depth-from-defocus

摘要

Depth-from-defocus (DFD) is useful in 3-D range image acquisition and automatic focusing. However, DFD needs precise calibration of internal camera parameters. Most DFD techniques use well known camera calibration techniques or specially designed vision systems to precisely measure the settings of the lens system of a camera. In this paper, we introduce an image-based defocus calibration technique. Especially we employ the Spatial domain Convolution/Deconvolution Transform Method (STM) which is introduced by Sub-barao and Surya [Subbarao, M., Surya, G., 1994. Depth-from-defocus: A spatial domain approach. Internat. J. Comput. Vision 13 (3), 271-294]. STM estimates blur parameters σ_1 and σ_2 of defocused images obtained at two different lens steps of the camera. In STM, the blur difference between two defocused images is considered to be a constant value which is determined by internal camera parameters. We use the blur level of the defocused images to calibrate blur difference and camera parameters. Calibrated parameters yield consistent results in depth estimation. Experimental results of depth measurement using real objects are presented.