Camera and 3D LiDAR sensors have become indispensable devices in modern autonomous driving vehicles. Camera provides fine-grained texture and color information in 2D space, while LiDAR captures more precise and farther-away distance measurements of the surrounding environments. The complementary information from these two sensors makes the fusion of two modalities a desired option. However, two primary challenges in the fusion of camera and LiDAR hinder its performance, i.e., how to effectively fuse the information from these two modalities and how to precisely align them (suffering from the weak spatiotemporal synchronization problem). This paper proposes a coarse-to-fine LiDAR and camera fusion-based network, named LIF-Seg, for LiDAR segmentation. For the first challenge, unlike these previous works fusing the point cloud and image information in a one-to-one manner, the proposed method introduces a simple but effective early-fusion strategy to fully utilize the contextual information of images. Second, to tackle the weak spatiotemporal synchronization problem, an offset rectification approach is designed to align the features of the two modalities. The cooperation of these two components leads to the success of the effective camera-LiDAR fusion. Experimental results on the nuScenes dataset show the superiority of LIF-Seg over existing methods by a large margin. Ablation studies and analyses further illustrate that the LIF-Seg can effectively address the weak spatiotemporal synchronization problem.

In this paper, we propose some efficient multi-view stereo methods for accurate and complete depth map estimation. We first present our basic methods with Adaptive Checkerboard sampling and Multi-Hypothesis joint view selection (ACMH & ACMH+). Based on our basic models, we develop two frameworks to deal with the depth estimation of ambiguous regions (especially low-textured areas) from two different perspectives: multi-scale information fusion and planar geometric clue assistance. For the former one, we propose a multi-scale geometric consistency guidance framework (ACMM) to obtain the reliable depth estimates for low-textured areas at coarser scales and guarantee that they can be propagated to finer scales. For the latter one, we propose a planar prior assisted framework (ACMP). We utilize a probabilistic graphical model to contribute a novel multi-view aggregated matching cost. At last, by taking advantage of the above frameworks, we further design a multi-scale geometric consistency guided and planar prior assisted multi-view stereo (ACMMP). This greatly enhances the discrimination of ambiguous regions and helps their depth sensing. Experiments on extensive datasets show our methods achieve state-of-the-art performance, recovering the depth estimation not only in low-textured areas but also in details. Related codes are available at https://github.com/GhiXu.

The traditional visual SLAM systems take the monocular or stereo camera as input sensor, with complex map initialization and map point triangulation steps needed for 3D map reconstruction, which are easy to fail, computationally complex and can cause noisy measurements. The emergence of RGB-D camera which provides RGB image together with depth information breaks this situation. While a number of RGB-D SLAM systems have been proposed in recent years, the current classification research on RGB-D SLAM is very lacking, and their advantages and shortcomings remain unclear regarding different applications and perturbations, such as illumination transformation, noise and rolling shutter effect of sensors. In this paper, we mainly introduced the basic concept and structure of the RGB-D SLAM system, and then introduced the differences between the various RGB-D SLAM systems in the three aspects of tracking, mapping, and loop detection, and we make a classification study on different RGB-D SLAM algorithms according to the three aspect. Furthermore, we discuss some advanced topics and open problems of RGB-D SLAM, hoping that it will help for future exploring. In the end, we conducted a large number of evaluation experiments on multiple RGB-D SLAM systems, and analyzed their advantages and disadvantages, as well as performance differences in different application scenarios, and provided references for researchers and developers.

The main challenges of visual object tracking arise from the arbitrary appearance of the objects that need to be tracked. Most existing algorithms try to solve this problem by training a new model to regenerate or classify each tracked object. As a result, the model needs to be initialized and retrained for each new object. In this paper, we propose to track different objects in an object-independent approach with a novel two-flow convolutional neural network (named YCNN). The YCNN takes two inputs (one is an object image patch, the other is a larger searching image patch), then outputs a response map which predicts how likely and where the object would appear in the search patch. Unlike the object-specific approaches, the YCNN is actually trained to measure the similarity between the two image patches. Thus, this model will not be limited to any specific object. Furthermore, the network is end-to-end trained to extract both shallow and deep dedicated convolutional features for visual tracking. And once properly trained, the YCNN can be used to track all kinds of objects without further training and updating. As a result, our algorithm is able to run at a very high speed of 45 frames-per-second. The effectiveness of the proposed algorithm can also be proved by the experiments on two popular datasets: OTB-100 and VOT-2014.

3D LiDAR has become an indispensable sensor in autonomous driving vehicles. In LiDAR-based 3D point cloud semantic segmentation, most voxel-based 3D segmentors cannot efficiently capture large amounts of context information, resulting in limited receptive fields and limiting their performance. To address this problem, a sparse voxel-based attention network is introduced for 3D LiDAR point cloud semantic segmentation, termed SVASeg, which captures large amounts of context information between voxels through sparse voxel-based multi-head attention (SMHA). The traditional multi-head attention cannot directly be applied to the non-empty sparse voxels. To this end, a hash table is built according to the incrementation of voxel coordinates to lookup the non-empty neighboring voxels of each sparse voxel. Then, the sparse voxels are grouped into different groups, and each group corresponds to a local region. Afterwards, position embedding, multi-head attention and feature fusion are performed for each group to capture and aggregate the context information. Based on the SMHA module, the SVASeg can directly operate on the non-empty voxels, maintaining a comparable computational overhead to the convolutional method. Extensive experimental results on the SemanticKITTI and nuScenes datasets show the superiority of SVASeg.

Outlier removal is a critical part of feature-based point cloud registration. In this paper, we revisit the model generation and selection of the classic RANSAC approach for fast and robust point cloud registration. For the model generation, we propose a second-order spatial compatibility (SC $^{2}$ ) measure to compute the similarity between correspondences. It takes into account global compatibility instead of local consistency, allowing for more distinctive clustering between inliers and outliers at an early stage. The proposed measure promises to find a certain number of outlier-free consensus sets using fewer samplings, making the model generation more efficient. For the model selection, we propose a new Feature and Spatial consistency constrained Truncated Chamfer Distance (FS-TCD) metric for evaluating the generated models. It considers the alignment quality, the feature matching properness, and the spatial consistency constraint simultaneously, enabling the correct model to be selected even when the inlier rate of the putative correspondence set is extremely low. Extensive experiments are carried out to investigate the performance of our method. In addition, we also experimentally prove that the proposed SC $^{2}$ measure and the FS-TCD metric are general and can be easily plugged into deep learning based frameworks. The code will be available at https://github.com/ZhiChen902/SC2-PCR-plusplus .