A Novel Image Representation via Local Frequency Analysis for Illumination Invariant Stereo Matching

Abstract

In this paper, we propose a novel image representation approach to tackle illumination variations in stereo matching problems. Images are mapped using their Fourier transforms which are convolved with a set of monogenic filters. Frequency analysis is carried out at different scales to account for most image content. The phase congruency and the local weighted mean phase angle are then computed over all the scales. The original image is transformed into a new representation using these two mappings. This representation is invariant to illumination and contrast variations. More importantly, it is generic and can be used with most sparse as well as dense stereo matching algorithms. In addition, sequential feature matching or tracking can also benefit from our approach in varying radiometric conditions. We demonstrate the improvements introduced with our image mappings on well-established data sets in the literature as well as on our own experimental scenarios that include high dynamic range imagery. The experiments include both dense and sparse stereo and sequential matching algorithms where the latter is considered in the very challenging visual odometry framework.

Thermal Stereo Odometry for UAVs

“…These figures also include the relative travelled errors obtained using each algorithm. The travelled errors at a given frame are calculated by (1), as shown at the bottom of this page, where P = [X, Y, Z ] T are the estimated camera poses and GT = [GT X , GT Y , GT Z ] T correspond to the ground truth. These travelled errors are plotted for the whole trajectory for each sequence.…”

“…We also plot the disparity map and 3D PCl from a pair of visible-band stereo images (Fig. 19b-19d) capturing the same scene using the vision system proposed in [1]. We used a specifically tuned version of the standard semiglobal block matching algorithm [49] to generate the disparity map ( Fig.…”

“…Here, we focus on VO and demonstrate the accuracy of the proposed system in day-time as well as night-time conditions. In contrast to visibleband cameras, infrared provide inherent robustness against illumination variations known to induce serious challenges to standard cameras [1]. Thermal sensors capture temperature variations within the scene and hence can be used in low-light and dark environments without requiring additional lighting.…”

Image Matching Using Enhancement Offsets With Adaptive Parameter Selection Via Histogram Analysis

“…Non-illumination significantly impacts matching performance [12,13]. The designed illumination offset is able to correct for simultaneous under or overexposure in an image.…”

A Survey on Motion Estimation Techniques

“…Motion estimation is applicable for video compression and enhancement [20], stereo correspondence [21], object tracking [22] and determination of optical flow [23]. Many specific methods have been proposed in the context of one or more of these applications, most methods require a matching metric that measures how similar two pixels or two patches of pixels are and hence how likely it is that they belong to the same surface point in the scene.…”