Keyframe Matching Research Articles

Feature- and Distribution-Based LiDAR SLAM With Generalized Feature Representation and Heuristic Nonlinear Optimization

In this article, we propose a novel 3-D light detection and ranging (LiDAR) simultaneous localization and mapping (SLAM) method, which contains feature-based fast scan matching, distribution-based keyframe matching, and loop closure. First, by incorporating feature-based and distribution-based LiDAR scan matching into our SLAM system, we combine the advantages of the two mechanisms and formulate a more efficient and precise front end. Second, considering the local surface reconstructing errors caused by the existing feature representation mechanisms, we propose a generalized representation mechanism for geometry feature representation within the distribution-based framework. Third, due to the existence of noisy point correspondences during the pose estimation optimization process, we propose a novel heuristic knowledge to distinguish high-quality matching error items and put more weights on them during the iterative objective function construction process. Extensive experiments are conducted on public and custom datasets, including KITTI, Mulran, NCLT, Stevens, and USTC-VLP16 datasets collected by our mobile platform. The experimental results demonstrate that the proposed mechanisms can effectively promote the LiDAR SLAM system performance and demonstrate competitive performance compared with state-of-the-art methods. We will make our source code open to serve as a new baseline for the robotic community.

An improved ORB-SLAM2 algorithm based on image information entropy

Aiming at the feature point extraction process in visual ORB-SLAM2, a visual SLAM method based on information entropy is proposed. In the process of ORB-SLAM2, due to the uncertainty of illumination and surrounding environment information, when the environment information is reduced, the number of feature points is likely to decrease dramatically, which results in the failure of SLAM process. In order to solve the above problems, a visual SLAM method based on image information entropy is proposed. According to the information entropy of the image, the amount of information is determined, and the image with low contrast and small gradient change is screened. The image is enhanced, and the feature points that can represent the image information are extracted as much as possible as the correlation basis of adjacent frame matching and key frame matching. In order to improve the performance of the mobile robot system, the local and global optimization of cumulative error of position and attitude is carried out by combining the optimization method of attitude graph.

A Sparse Visual Odometry Technique Based on Pose Adjustment With Keyframe Matching

In this article we propose a sparse visual odometry model for RGB-D images. The technique utilizes the minimization of the photometric errors obtained from the edge features for pose adjustment. Different from the conventional feature-based approaches, the features are extracted on the edge images. It makes the feature matching more robust and the computation more efficient. Moreover, we introduce a posterior probability based on the different degree of exposure for each edge point. The weight is then adjusted to improve the pose estimation by keyframe matching according to the probability. Since the sensor noise will affect the feature extraction results and cause the inaccurate estimation, the epipolar geometry and mixture distribution are used for the depth value updates. The experiments carried out using public datasets and our own image sequences have demonstrated the effectiveness of the proposed technique.

Image-Based Localization Using Context

<p>Image-based localization problem consists of estimating the 6 DoF<br />camera pose by matching the image to a 3D point cloud (or equivalent)<br />representing a 3D environment. The robustness and accuracy<br />of current solutions is not objective and quantifiable. We<br />have completed a comparative analysis of the main state of the art<br />approaches, namely Brute Force Matching, Approximate Nearest<br />Neighbour Matching, Embedded Ferns Classification, ACG Localizer(<br />Using Visual Vocabulary) and Keyframe Matching Approach.<br />The results of the study revealed major deficiencies in each approach<br />mainly in search space reduction, clustering, feature matching<br />and sensitivity to where the query image was taken. Then, we<br />choose to focus on one common major problem that is reducing<br />the search space. We propose to create a new image-based localization<br />approach based on reducing the search space by using<br />global descriptors to find candidate keyframes in the database then<br />search against the 3D points that are only seen from these candidates<br />using local descriptors stored in a 3D cloud map.</p>

Motion Redirection Based on Markov Chain Monte Carlo Particle Filter

This paper aims at the difficulty that lack of observation model and high-dimensional sampling in video tooning, proposes a method based on key frame matching and dual-directional Markov chain Monte Carlo sampling of video motion redirection. At first, after extracting the key frame of a given video, By affine transformation and linear superposition, the subject initializes the video’s space-time parameters and forms the observation model; Secondly, in each space-time, based on the bi-directional Markov property of each frame, This paper proposed a dual-directional Markov chain Monte Carlo sampling particle filter structure and takes full advantage of the relationship of the front and back frame of the parameters to estimate motion redirection parameters. At the same time, for high-dimensional sampling problem, the subject according to the directional parameters’ correlation implements classification of skeleton parameters-morphological parameters-physical parameters, proposes a hierarchical genetic strategy to optimize the output parameters and improves the efficiency of the algorithm. The research of this paper will produce an efficient and prominent animation expressive video motion redirection method and play an important role on video animation of the development.

Automated sampling and control of gaseous simulations

In this work, we describe a method that automates the sampling and control of gaseous fluid simulations. Several recent approaches have provided techniques for artists to generate high-resolution simulations based on a low-resolution simulation. However, often in applications the overall flow in the low-resolution simulation that an animator observes and intends to preserve is composed of even lower frequencies than the low resolution itself. In such cases, attempting to match the low-resolution simulation precisely is unnecessarily restrictive. We propose a new sampling technique to efficiently capture the overall flow of a fluid simulation, at the scale of user's choice, in such a way that the sampled information is sufficient to represent what is virtually perceived and no more. Thus, by applying control based on the sampled data, we ensure that in the resulting high-resolution simulation, the overall flow is matched to the low-resolution simulation and the fine details on the high resolution are preserved. The samples we obtain have both spatial and temporal continuity that allows smooth keyframe matching and direct manipulation of visible elements such as smoke density through temporal blending of samples. We demonstrate that a user can easily configure a simulation with our system to achieve desired results.

Efficient subsequence matching over large video databases

Video similarity matching has broad applications such as copyright detection, news tracking and commercial monitoring, etc. Among these applications, one typical task is to detect the local similarity between two videos without the knowledge on positions and lengths of each matched subclip pair. However, most studies so far on video detection investigate the global similarity between two short clips using a pre-defined distance function. Although there are a few works on video subsequence detection, all these proposals fail to provide an effective query processing mechanism. In this paper, we first generalize the problem of video similarity matching. Then, a novel solution called consistent keyframe matching (CKM) is proposed to solve the problem of subsequence matching based on video segmentation. CKM is designed with two goals: (1) good scalability in terms of the query sequence length and the size of video database and (2) fast video subsequence matching in terms of processing time. Good scalability is achieved by employing a batch query paradigm, where keyframes sharing the same query space are summarized and ordered. As such, the redundancy of data access is eliminated, leading to much faster video query processing. Fast subsequence matching is achieved by comparing the keyframes of different video sequences. Specifically, a keyframe matching graph is first constructed and then divided into matched candidate subgraphs. We have evaluated our proposed approach over a very large real video database. Extensive experiments demonstrate the effectiveness and efficiency of our approach.