Denoising techniques for raw 3D data of TOF cameras based on clustering and wavelets

Abstract

In order to measure the 3D structure of a number of objects a comparably new technique in computer vision exists, namely time of flight (TOF) cameras. The overall principle is rather easy and has been applied using sound or light for a long time in all kind of sonar and lidar systems. However in this approach one uses modulated light waves and receives the signals by a parallel pixel array structure. Out of the travelling time at each pixel one can estimate the depth structure of a distant object. The technique requires measuring the intensity differences and ratios of several pictures with extremely high accuracy; therefore one faces in practice rather high noise levels. Object features as reflectance and roughness influence the measurement results. This leads to partly high noise levels with variances dependent on the illumination and material parameters. It can be shown that a reciprocal relation between the variance of the phase and the squared amplitude of the signals exists. On the other hand, objects can be distinguished using these dependencies on surface characteristics. It is shown that based on local variances assigned to separated objects appropriate denoising can be performed based on Wavelets and edge-preserving smoothing methods.

A Clustering Based Denoising Technique for Range Images of Time of Flight Cameras

“…The identified clusters are used to estimate the noise level inside each cluster, and smoothing is then performed on each cluster tuned to its characteristics. The details of this approach are given in [8]; the following gives an overview of this approach.…”

“…Additionally, intensity, location and noise characteristics are incorporated. Details are given in [8]. The actual smoothing uses a Gaussian kernel.…”

Denoising of Time-of-Flight depth data via iteratively reweighted least squares minimization

“…To preserve depth discontinuities, a line of approaches [9,10] applies clustering techniques in multi-dimensional feature spaces formed by the measured distance and amplitude values. Another line of approaches [8,11] uses non-local means filters, which preserve the structural information of the scene using the patch-wise depth and amplitude similarity between pixels.…”

Denoising Time-Of-Flight Data with Adaptive Total Variation

“…Denoising approaches for ToF data are e.g. presented in [15,10,7,6,11,1]. These approaches are mainly based on bilateral filtering or wavelet techniques, combined with appropriate noise models.…”

“…In order to regularize the denoising problem, i.e. prescribing smoothness of the result, these models (except for the clustering approach in [15]) do not assume explicit geometric structures in the depth map such as regular edges or piecewise planar surfaces. In contrast, we present here a variational denoising approach based on adaptive total variation (TV), which allows to take into account such geometric properties and thus to reconstruct edges and slopes with sufficient regularity.…”