This paper was published in Proc. SPIE Vol. 6762 and is made available as an electronic reprint with permission of SPIE. One print or electronic copy may be made for personal use only. Systematic or multiple reproduction, distribution to multiple locations via electronic or other means, duplication of any material in this paper for a fee or for commercial purposes, or modification of the content of the paper are prohibited. Optimized scattering compensation for time-of-flight camera James Mure-Dubois and Heinz H¨ ugli University of Neuchˆatel - Institute of Microtechnology, 2000 Neuchˆatel, Switzerland ABSTRACT Recent time-of-flight (TOF) cameras allow for real-time acquisition of range maps with good performance. However, the accuracy of the measured range map may be limited by secondary light reflections. Specifically, the range measurement is affected by scattering, which consists in parasitic signals caused by multiple reflections inside the camera device. Scattering, which is particularly strong in scenes with large aspect ratios, must be detected and the errors compensated. This paper considers reducing scattering errors by means of image processing methods applied to the output image from the time-of-flight camera. It shows that scattering reduction can be expressed as a deconvolution problem on a complex, two-dimensional signal. The paper investigated several solutions. First, a comparison of image domain and Fourier domain processing for scattering compensation is provided. One key element in the comparison is the computation load and the requirement to perform scattering compensation in real-time. Then, the paper discusses strategies for improved scattering reduction. More specifically, it treats the problem of optimizing the description of the inverse filter for best scattering compensation results. Finally, the validity of the proposed scattering reduction method is verified on various examples of indoor scenes. Keywords: 3D vision, time-of-flight, range imaging, scattering, deconvolution, range camera 1. INTRODUCTION Recent time-of-flight (TOF) cameras allow for real-time acquisition of range maps with good performance. For instance, the Swissranger SR-3000 camera has a 176 × 144 pixels sensor, and supports continuous operation at 20 Hz. 1 The depth resolution can be better than 1 cm in favourable conditions, i.e. with no bright light sources in the field of view. 2 However, depth measurement is degraded by scattering, which consists in secondary reflections occurring between the lens and the imager 3, 4, . 5 The degradation is particularly strong when the weak signal from far objects (background) is affected by the strong scattering signal from foreground objects. This degradation of the depth image is a significant penalty in many applications, especially when background subtraction methods are employed. 3 For this reason, scattering must be suppressed, or at least reduced. In [5], scattering reduction was expressed as a deconvolution problem on a complex, two-dimensional signal. The proposed algorithm for scattering reduction uses inverse filters implemented by a sum of separable gaussians. In this paper, we will discuss two improvements to this approach. The first improvement is the usage of Fourier domain processing for convolution operations. Processing in Fourier domain brings the advantage that the computation time per frame is significantly shorter than for spatial filtering. The second contribution concerns the degradation modeling. Given that an accurate measurement of the PSF is not possible, the paper considers various possible descriptions for the scattering. Finally, parametric models where the scattering degradation is expressed as a sum of gaussian filters are used. A method for parameter estimation is described that opens to the automatic determination of a camera specific scattering compensation model. The remainder of this paper is organized as follows. A brief reminder of time-of-flight imaging principles is presented in section 2. Section 3 discusses image domain versus Fourier domain processing. The different descriptions for usable scattering models are discussed in section 4. Finally, section 5 presents real-world results illustrating the performance of scattering compensation. Copyright 2007 Society of Photo-Optical Instrumentation Engineers Proc SPIE Vol. 6762 - 67620H-1