Enhancement of light field disparity maps by reducing the silhouette effect and plane noise

During the last decade, there has been an increasing number of applications dealing with multidimensional visual information, either for 3D object representation or feature extraction purposes. In this context, recent advances in light field technology, have been driving research efforts in disparity estimation methods. Among the existing ones, those based on the structure tensor have emerged as very promising to estimate disparity maps from Epipolar Plane Images. However, this approach is known to have two intrinsic limitations: (i) silhouette enlargement and (ii) irregularity of surface normal maps as computed from the estimated disparity. To address these problems, this work proposes a new method for improving disparity maps obtained from the structure-tensor approach by enhancing the silhouette and reducing the noise of planar surfaces in light fields. An edge-based approach is initially used for silhouette improvement through refinement of the estimated disparity values around object edges. Then, a plane detection algorithm, based on a seed growth strategy, is used to estimate planar regions, which in turn are used to guide correction of erroneous disparity values detected in object boundaries. The proposed algorithm shows an average improvement of 98.3% in terms of median angle error for plane surfaces, when compared to regular structure-tensor-based methods, outperforming state-of-the-art methods. The proposed framework also presents very competitive results, in terms of mean square error between disparity maps and their ground truth, when compared with their counterparts.

Disparity Estimation of High-Resolution Remote Sensing Images with Dual-Scale Matching Network

As an essential task in remote sensing, disparity estimation of high-resolution stereo images is still confronted with intractable problems due to extremely complex scenes and dynamically changing disparities. Especially in areas containing texture-less regions, repetitive patterns, disparity discontinuities, and occlusions, stereo matching is difficult. Recently, convolutional neural networks have provided a new paradigm for disparity estimation, but it is difficult for current models to consider both accuracy and speed. This paper proposes a novel end-to-end network to overcome the aforementioned obstacles. The proposed network learns stereo matching at dual scales, in which the low one captures coarse-grained information while the high one captures fine-grained information, helpful for matching structures of different scales. Moreiver, we construct cost volumes from negative to positive values to make the network work well for both negative and nonnegative disparities since the disparity varies dramatically in remote sensing stereo images. A 3D encoder-decoder module formed by factorized 3D convolutions is introduced to adaptively learn cost aggregation, which is of high efficiency and able to alleviate the edge-fattening issue at disparity discontinuities and approximate the matching of occlusions. Besides, we use a refinement module that brings in shallow features as guidance to attain high-quality full-resolution disparity maps. The proposed network is compared with several typical models. Experimental results on a challenging dataset demonstrate that our network shows powerful learning and generalization abilities. It achieves convincing performance on both accuracy and efficiency, and improvements of stereo matching in these challenging areas are noteworthy.

Three-Dimensional Reconstruction of Light Field Based on Phase Similarity

Light field imaging plays an increasingly important role in the field of three-dimensional (3D) reconstruction because of its ability to quickly obtain four-dimensional information (angle and space) of the scene. In this paper, a 3D reconstruction method of light field based on phase similarity is proposed to increase the accuracy of depth estimation and the scope of applicability of epipolar plane image (EPI). The calibration method of the light field camera was used to obtain the relationship between disparity and depth, and the projector calibration was removed to make the experimental procedure more flexible. Then, the disparity estimation algorithm based on phase similarity was designed to effectively improve the reliability and accuracy of disparity calculation, in which the phase information was used instead of the structure tensor, and the morphological processing method was used to denoise and optimize the disparity map. Finally, 3D reconstruction of the light field was realized by combining disparity information with the calibrated relationship. The experimental results showed that the reconstruction standard deviation of the two objects was 0.3179 mm and 0.3865 mm compared with the ground truth of the measured objects, respectively. Compared with the traditional EPI method, our method can not only make EPI perform well in a single scene or blurred texture situations but also maintain good reconstruction accuracy.

Efficient and Accurate Disparity Estimation from MLA-Based Plenoptic Cameras

Microlens-array based plenoptic cameras capture the light field in a single shot, enabling new potential applications but also introducing additional challenges. A plenoptic image consists of thousand of microlens images. Estimating the disparity for each microlens allows to render conventional images, changing the perspective and the focal settings, and to reconstruct the three-dimensional geometry of the scene. The work includes a blur-aware calibration method to model plenoptic cameras, an optimization method to accurately select the best microlenses combination for disparity estimation, an overview of the different types of plenoptic cameras, an analysis of the disparity estimation algorithms, and a robust depth estimation approach for light field microscopy. The research led to the creation of a full framework for plenoptic cameras, which contains the implementation of the algorithms discussed in the work and datasets of both real and synthetic images for comparison, benchmarking and future research.

A Fast Stereo Matching Network with Multi-Cross Attention

Stereo matching networks based on deep learning are widely developed and can obtain excellent disparity estimation. We present a new end-to-end fast deep learning stereo matching network in this work that aims to determine the corresponding disparity from two stereo image pairs. We extract the characteristics of the low-resolution feature images using the stacked hourglass structure feature extractor and build a multi-level detailed cost volume. We also use the edge of the left image to guide disparity optimization and sub-sample with the low-resolution data, ensuring excellent accuracy and speed at the same time. Furthermore, we design a multi-cross attention model for binocular stereo matching to improve the matching accuracy and achieve end-to-end disparity regression effectively. We evaluate our network on Scene Flow, KITTI2012, and KITTI2015 datasets, and the experimental results show that the speed and accuracy of our method are excellent.