This paper presents a novel approach for vehicle map mapping and estimating parking occupancy in parking lots using a single camera without parking lines. The system is capable of detecting around 10 vehicles even with just one camera, and is able to track vehicles with the same ID as they enter and pass through the parking lot. Furthermore, it proposes a method to map the locations of vehicles on real-world maps, allowing for accurate estimation of parking occupancy. The system achieved high accuracy in vehicle detection and tracking, with a 98.4% recall rate and 100% tracking accuracy. The accuracy of parking occupancy estimation was calculated using root mean square error, with an average error of 0.24m and a maximum error of 0.36m. The results demonstrate the feasibility of detecting and tracking vehicles, as well as estimating parking occupancy, in parking lots without parking lines using only one camera. Further testing on multiple sites, including nighttime and adverse weather conditions, is needed to increase the reliability of the system.

Multiple motion estimators such as an interacting multiple model (IMM) have been utilized to track target objects such as cars and pedestrians with diverse motion patterns. However, the standard IMM has limitations in combining motion models with different state definitions, so it cannot contain a complementary set of models that accurately work for all motion patterns. In this paper, we propose IMM-based adaptive target tracking with heterogeneous velocity representations and linear/curvilinear motion models. It can integrate four motion models with different state definitions and dimensions to be completely complimentary for all types of motions. We experimentally demonstrate the effectiveness of the proposed method with accuracy for various motion patterns using two types of datasets: synthetic datasets and real datasets. Experimental results show that the proposed method achieves the adaptive target tracking for diverse types of motion and also for various objects such as cars, pedestrians, and drones in the real world.

A GPS is usually installed at the center of a vehicle because with this configuration its data become the position of the vehicle without complex calibration and transformation. In this paper, however, we verify that the off-center arrangement of the GPS improves the position and orientation accuracy theoretically and experimentally. The extended Kalman filter (EKF) formulation and its observability and uncertainty analysis present that the off-centered GPS is better observable and less uncertain. Moreover, our experiments with synthetic data and real sensor dataset confirm the improvement of localization accuracy. The synthetic data generated in various situations reveal the important properties of localization with the off-centered GPS. The public outdoor dataset with real GPS data also supports the advantage of the off-centered GPS.

In this paper, we present a method that is able to detect abnormal behavior of intoxicated people in surveillance videos. We first describe typical behavior patterns of intoxicated people in videos and derive two visual features that distinguish them effectively. We define a motion efficiency as one feature to capture intoxicated motion and the aspect ratio of a bounding box of an object as the other to detect intoxicated postures. For the computation of the proposed visual features, the method detects and tracks individual pedestrians in videos and evaluates their motion trajectories and pose trajectories, respectively. The experimental results on the test dataset on railway platform show that the proposed method is able to detect drunken passengers effectively and robustly in a real environment.

This paper introduces about an open-source project, Triangulation Toolbox, for landmark-based localization. This project aims to share various algorithms and evaluate their performance in public. At first, landmark-based localization is briefly reviewed with respect to its types of input measurements. From its generalized problem formulation, we adopt a common interface for algorithms in the toolbox. Each algorithm is based on diverse types of input measurements (e.g. distance or bearing angle) in different dimension of space (e.g. 2D or 3D). To support capricious requirements, the toolbox follows the most general convention on measurements and space. In our performance evaluation, nine algorithms were experimentally compared in the view of position and orientation accuracy. The evaluation also leads helpful comments on landmark-based localization and its applications. Since Triangulation Toolbox is designed to be simple and flexible, many developers and researchers can refer implemented algorithms for their applications and also evaluate their own algorithm through the benchmark.

3D range sensors are currently being used in various fields. Creating 3D range sensors requires various techniques, such as object detection, tracking, classification, 3D SLAM, etc. For the pre-processing step, superpixels can improve the performance of these techniques. This paper proposes a novel over-segmentation algorithm, known as superpixels, for 3D outdoor urban range data. Superpixels are generated with three steps: boundary extraction using a surface change score and sensor models, initial cluster seeding using a quadtree decomposition, and iterative clustering, which adapts a k-means clustering approach with limited search size in the quadtree dimension. The proposed algorithm produces adaptive superpixel sizes that take into account surface and object border information. This reduces memory size more than regular grid methods and represents small objects well with adaptable pixel sizes. The algorithm is verified using the publicly available Velodyne dataset and the manually annotated ground truth. A comparison with the conventional algorithm is also presented.

This paper presents a novel object detection and segmentation method utilizing an inpainting algorithm. Inpainting is a concept of recovering missing image regions based on their surroundings, which were originally used for restoration of damaged paintings. In this paper, we newly utilize inpainting to judge whether an object candidate region includes the foreground object or not. The key idea is that if we erase a certain region from an image, the inpainting algorithm is expected to recover the erased image only when it belongs a background area (i.e. only when there is no object in it). By measuring the similarity between the inpainted region and the original image region, our approach filters out false detections while maintaining true object detections. Furthermore, we take advantage of the inpainting for object segmentation, since our approach is designed to explicitly distinguish foreground areas from its background. Experimental results confirm that our approach applied to baseline detectors enables better recognition of objects, obtaining higher accuracies. We illustrate how our inpainting-based detection/segmentation approach benefits the object detection using two different pedestrian datasets.

A* on grid maps generates a path with zig-zag pattern, but Theta* is known to be free from this disadvantage. Theta* assumes that cost of each cell, cell-cost, is uniform, but non-uniform costs are effective ways to represent traversability on grid maps. Theta* does not work on non-uniform costmaps. In this paper, we generalize Theta* toward non-uniform costmaps. To extend Theta*, we propose two kinds of cost functions considering non-uniform cell-costs. The first function adopts the arithmetic mean under the assumption that all cells contribute equally to the overall cost. The second function uses the weighted mean by considering the true traversal length on each cell. We applied the proposed methods to two types of maps: synthetic and real maps. An experiment on synthetic maps quantifies performance of the two methods in terms of accuracy and computing time. The other experiment on real maps presents the effectiveness of Theta* with the proposed methods. The generalized Theta* generated the least-cost path compared with the original Theta* and A* on non-uniform costmaps.

In this paper, we introduce the concept of personal driving diary. A personal driving diary is a multimedia archive of a person's daily driving experience, describing important driving events of the user with annotated videos. This paper presents an automated system that constructs such multimedia diary by analyzing videos obtained from a vehicle-mounted camera. The proposed system recognizes important interactions between the driving vehicle and the others from videos (e.g. accident, overtaking, ...), and labels them together with its contextual knowledge on the vehicle (e.g. its physical location on the map) to construct an event log. A novel decision tree based activity recognizer that incrementally learns driving events from first-person view videos is designed. The constructed diary enables efficient searching and event-based browsing of video clips, which helps the user to retrieve videos of dangerous situations and analyze his/her driving habits statistically. Our experiment confirms that the proposed system reliably generates driving diaries by annotating learned vehicle events.

Although GPS/DGPS become the dominant localization solution in the outdoor environment, it needs assistant sensors or algorithms for the covering the area not to get the position information from GPS. Especially, in the robot navigation, the sensor fusion algorithm is needed. In addition, it is hard to get the position information at the area surrounded the high buildings such as the downtown because GPS signals is so feeble. Therefore, this paper illustrates an efficient method for the outdoor localization incorporating DGPS, Encoder, and IMU sensor based on EKF. To show the localization performances of the proposed fusion algorithm, we have implemented the proposed algorithm and applied the advertising robot platform which is operating well during 80 days in the real semi-outdoor structured environment. The proposed sensor fusion algorithm and the experimental results showed the feasibility of our novel sensor fusion algorithm.

RANSAC (Random Sample Consensus) has been popular in regression problem with samples contaminated with outliers. It has been a milestone of many researches on robust estimators, but there are a few survey and performance analysis on them. This paper categorizes them on their objectives: being accurate, being fast, and being robust. Performance evaluation performed on line fitting with various data distribution. Planar homography estimation was utilized to present performance in real data.

The core step of video stabilization is to estimate global motion from locally extracted motion clues. Outlier motion clues are generated from moving objects in image sequence, which cause incorrect global motion estimates. Random sample consensus (RANSAC) is popularly used to solve such outlier problem. RANSAC needs to tune parameters with respect to the given motion clues, so it sometimes fail when outlier clues are increased than before. Adaptive RANSAC is proposed to solve this problem, which is based on maximum likelihood sample consensus (MLESAC). It estimates the ratio of outliers through expectation maximization (EM), which entails the necessary number of iteration for each frame. The adaptation sustains high accuracy in varying ratio of outliers and faster than RANSAC when fewer iteration is enough. Performance of adaptive RANSAC is verified in experiments using four images sequences.

Data may be wrongly measured or come from other sources. Such data is a big problem in regression, which retrieve parameters from data. Random sample consensus (RANSAC) and maximum likelihood estimation sample consensus (MLE-SAC) are representative researches, which focused on this problem. However, they do not cope with varying data distribution because they need to tune variables according to given data. This paper proposes user-independent parameter estimator, u-MLESAC, which is based on MLESAC. It estimates variables necessary in probabilistic error model through expectation maximization (EM). It also terminates adaptively using failure rate and error tolerance, which can control trade-off between accuracy and running time. Line fitting experiments showed its high accuracy and robustness in varying data distribution. Its results are compared with other estimators. Its application to landmark-based localization also verified its performance compared with other estimator.

This study investigates the nonlinear dynamics of traction for an omniwheel mobile robot platform. A nonlinear slip model is incorporated into the dynamics of the system and the resulting equations of motion are derived using Euler-Lagrange formulation. These are additionally transformed into slip-space, where the dynamical equations lend themselves to a convenient analysis of the slip dynamics. The conventional assumptions for ideal rolling are also explored and a reduced expression for the nonlinear dynamics is generated for such situations. Preliminary explorations toward a comprehensive analysis of the dynamics for omniwheel platforms under various control schemes is also initiated.