Search

Detecting buildings in the surroundings of an urban vehicle and matching them to building models available on map services is an emerging trend in robotics localization for urban vehicles. In this paper, we present a novel technique, which improves a previous work by detecting building façade, their positions, and finding the correspondences with their 3D models, available in OpenStreetMap. The proposed technique uses segmented point clouds produced using stereo images, processed by a convolutional neural network. The point clouds of the façades are then matched against a reference point cloud, produced extruding the buildings’ outlines, which are available on OpenStreetMap (OSM). In order to produce a lane-level localization of the vehicle, the resulting information is then fed into our probabilistic framework, called Road Layout Estimation (RLE). We prove the effectiveness of this proposal, testing it on sequences from the well-known KITTI dataset and comparing the results concerning a basic RLE version without the proposed pipeline.

The estimation of the 3D trajectory, the class and the dimensions of a vehicle represents three relevant tasks for traffic monitoring. They are usually performed by separate sub-systems and only few existing algorithms cope with the three tasks at the same time. However, if these tasks are integrated, the trajectory estimation enforces the classification with temporal consistency, and at the same time, the estimation of the vehicle class and dimensions can be used to increase the trajectory estimate accuracy. In this work, we propose an algorithm to estimate the 3D trajectory, the class and the dimensions of vehicles simultaneously by means of a Backward-Simulation Particle Smoother whose state contains both continuous (vehicle pose and dimensions), and discrete (vehicle class) quantities. To integrate the class estimate in the Particle Smoother we model the class prediction as a Markov Chain. We performed experimental tests on both simulated and real datasets; they show that the pose and dimension estimation reaches centimeter-accuracy and the classification accuracy is higher than 95 %.

Background subtraction is the classical approach to differentiate moving objects in a scene from the static background when the camera is fixed. If the fixed camera assumption does not hold, a frame registration step is followed by the background subtraction. However, this registration step cannot perfectly compensate camera motion, thus errors like translations of pixels from their true registered position occur. In this paper, we overcome these errors with a simple, but effective background subtraction algorithm that combines Temporal and Spatio-Temporal approaches. The former models the temporal intensity distribution of each individual pixel. The latter classifies foreground and background pixels, taking into account the intensity distribution of each pixels' neighborhood. The experimental results show that our algorithm outperforms the state-of-the-art systems in the presence of jitter, in spite of its simplicity. © 2013 Springer-Verlag Berlin Heidelberg.

Simultaneous Localization and Mapping (SLAM) has received quite a lot of attention in the last decades because of its relevance for many applications centered on a mobile observer, such as service robotics and intelligent transportation systems. This paper focuses on the use of recursive Bayesian filtering, as implemented by the Extendend Kalman Filter (EKF), to face the Visual SLAM problem, i.e., when using data from visual sources. In Monocular SLAM, which uses a single camera as unique source of information, it is not possible to directly estimate the depth of a feature from a single image. To handle the severely non-normal distribution representing such uncertainty, inverse parametrizations were developed, capable to deal with such uncertainty and still relying on Gaussian variables. In the paper, after an introduction to EKF-SLAM, we provide a review of different inverse parametrizations, and we introduce a novel proposal, the Framed Inverse Depth (FID) parametrization, which, in terms of consistency, performs similarly to state of the art Monocular SLAM parametrizations, but at a reduced computational cost. All these parametrizations can be used in a stereo and multi camera setting too. An extensive analysis is presented for both Monocular and stereo SLAM, for a simulated environment widely used in the literature as well as on a widely used real dataset.

In the last years, the Monocular SLAM problem was widely studied, to allow the simultaneous reconstruction of the environment and the localization of the observer, by using a single camera. As for other SLAM problems, a frequently used feature for the representation of the world, is the 3D point. Differently from other SLAM problems, because of the perspective model of the camera, in Monocular SLAM, features cannot be completely perceived and initialized from a single measurement. To solve this issue, different parameterizations have been proposed in the literature, which try to solve also another problem in Monocular SLAM, i.e., the distortion of the Gaussian uncertainty in depth estimation that takes place because of the nonlinear measurement model. In this paper, we start from recent results in consistency analysis for these parameterizations to propose a novel approach to improve EKF-based Monocular SLAM even further. Our claims are sustained by an extended validation on simulated and real data.

Recent works have shown that it is possible to solve the Simultaneous Localization And Mapping problem using an Extended Kalman Filter and a single perspective camera. The principal drawback of these works is an inaccurate modeling of measurement uncertainties, which therefore causes inconsistencies in the filter estimations. A possible solution to proper uncertainty modeling is the Unified Inverse Depth parametrization. In this paper we propose the Inverse Scaling parametrization that still allows an un-delayed initialization of features, while reducing the number of needed parameters and simplifying the measurement model. This novel approach allows a better uncertainty modeling of both low and high parallax features and reduces the likelihood of inconsistencies. Experiments in simulation demonstrate that the use of the Inverse Scaling solution improves the performance of the monocular EKF SLAM filter when compared with the Unified Inverse Depth approach; experiment on real data confirm the applicability of the idea.

A trustable and accurate ground truth is a key requirement for benchmarking self-localization and mapping algorithms; on the other hand, collection of ground truth is a complex and daunting task, and its validation is a challenging issue. In this paper we propose two techniques for indoor ground truth collection, developed in the framework of the European project Rawseeds, which are mutually independent and also independent on the sensors onboard the robot. These techniques are based, respectively, on a network of fixed cameras, and on a network of fixed laser scanners. We show how these systems are implemented and deployed, and, most importantly, we evaluate their performance; moreover, we investigate the possible fusion of their outputs. © 2009 Springer Science+Business Media, LLC.

The recent literature has shown that it is possible to solve the monocular Simultaneous Localization And Mapping using both undelayed features ini- tialization and an Extedend Kalman Filter. The key concept, to achieve this result, was the introduction of a new parametrization called Unified Inverse Depth that produces measurements equations with a high degree of linear- ity and allows an efficient and accurate modeling of uncertainties. In this paper we present a monocular EKF SLAM filter based on an alternative parametrization, i.e., the Inverse Scaling Parametrization, characterized by a reduced number of parameters, a more linear measurement model, and a bet- ter modeling of features uncertainty for both low and high parallax features. Experiments in simulation demonstrate that the use of the Inverse Scaling solution improves the monocular EKF SLAM filter when compared with the Unified Inverse Depth approach, while experiments on real data show the system working in practice as well.

Self localization and mapping with vision is still an open research field. Since redundancy in the sensing suite is too expensive for consumer-level robots, we base on vision as the main sensing system for SLAM. We approach the problem with 3D data from a trinocular vision system. Past experience shows that problems arise as a consequence of inaccurate modeling of uncertainties; interestingly enough, we found that accuracy in modeling the robot pose uncertainty is much less relevant than for the uncertainty on the sensed data. To overcome the severe limitation of linear and Gaussian approximations, we applied a particle-based description of the inherently non-normal probability density distribution of the sensed data; the aim is to increase the success rate of data association, which we see as the most important problem. The increase in correct data associations reduces the uncertainty in the model and, consequently, in the robot pose, respectively estimated with a hierarchical map decomposition and a six degree of freedom extended Kalman filter. In this paper, we present approaches for particle-based sensor modeling and data association, with a comparative experimental evaluation on real 3D vision data.

A method is proposed for calibrating Catadioptric Omni-directional Vision Systems. This method, similarly to classic camera calibration, makes use of a set of fiducial points to find the parameters of the geometric image formation model. The method makes no particular assumption regarding the shape of the mirror or its position with respect to the camera. Given the camera intrinsic parameters and the mirror profile, the mirror pose is computed using the projection of the mirror border and, eventually, the extrinsic parameters are computed by minimizing distance between fiducial points and back-projected images.

In this paper we propose an approach to color classification and image segmentation in non-stationary environments. Our goal is to cope with changing illumination condition by on-line adapting both the parametric color model and its structure/complexity. Other authors used parametric statistics to model color distribution in segmentation and tracking problems, but with a fixed complexity model. Our approach is able to on-line adapt also the complexity of the model, to cope with large variations in the scene illumination and color temperature.