Alejandro Rituerto

Visual SLAM with an Omnidirectional Camera

In this work we integrate the Spherical Camera Model for catadioptric systems in a Visual-SLAM application. The Spherical Camera Model is a projection model that unifies central catadioptric and conventional cameras. To integrate this model into the Extended Kalman Filter-based SLAM we require to linearize the direct and the inverse projection. We have performed an initial experimentation with omni directional and conventional real sequences including challenging trajectories. The results confirm that the omni directional camera gives much better orientation accuracy improving the estimated camera trajectory.

Semantic labeling for indoor topological mapping using a wearable catadioptric system

An important part of current research on appearance based mapping goes towards richer semantic representations of the environment, which may allow autonomous systems to perform higher level tasks and provide better human-robot interaction. This work presents a new omnidirectional vision based scene labeling approach for augmented indoor topological mapping. Omnidirectional vision systems are of particular interest because they allow us to have more compact and efficient representation of the environment. Our proposal includes novel ideas in order to augment the semantic information of a typical indoor topological map: we pay special attention to the semantic labels of the different types of transitions between places, and propose a simple way to include this semantic information to build a topological map, as part of the criteria to segment the environment. This work is built on efficient catadioptric image representation based on the Gist descriptor, which is used to classify the acquired views into types of indoor regions. The basic types of indoor regions considered are Place and Transition, farthest divided into more specific subclasses, e.g., Transition into door, stairs and elevator. Besides using the result of this labeling, the proposed mapping approach includes a probabilistic model to account for spatio-temporal consistency. All the proposed ideas have been evaluated in a new indoor dataset presented in this paper. This dataset has been acquired with our wearable catadioptric vision system 1 , showing promising results

Adapting a real-time monocular visual SLAM from conventional to omnidirectional cameras

The SLAM (Simultaneous Localization and Mapping) problem is one of the essential challenges for the current robotics. Our main objective in this work is to develop a real-time visual SLAM system using monocular omnidirectional vision. Our approach is based on the Extended Kalman Filter (EKF). We use the Spherical Camera Model to obtain geometric information from the images. This model is integrated in the EKF-based SLAM through the linearization of the direct and the inverse projections. We introduce a new computation of the descriptor patch for catadioptric omnidirectional cameras which aims to reach rotation and scale invariance. We perform experiments with omnidirectional images comparing this new approach with the conventional one. The experimentation confirms that our approach works better with omnidirectional cameras since features last longer and constructed maps are bigger

Wearable omnidirectional vision system for personal localization and guidance

Autonomous navigation and recognition of the environment are fundamental abilities for people extensively studied in computer vision and robotics fields. Expansion of low cost wearable sensing provides interesting opportunities for assistance systems that augment people navigation and recognition capabilities. This work presents our wearable omnidirectional vision system and a novel two-phase localization approach running on it. It runs state-of-the-art real time visual odometry adapted to catadioptric images augmented with topological-semantic information. The presented approach benefits from using wearable sensors to improve visual odometry results with true scaled solution. The wide field of view of catadioptric vision system used makes features last longer in the field of view and allows more compact location representation which facilitates topological place recognition. Experiments in this paper show promising ego-localization results in realistic settings, providing good true scaled visual odometry estimation and recognition of indoor regions.

Line image signature for scene understanding with a wearable vision system

Wearable computer vision systems provide plenty of opportunities to develop human assistive devices. This work contributes on visual scene understanding techniques using a helmet-mounted omnidirectional vision system. The goal is to extract semantic information of the environment, such as the type of environment being traversed or the basic 3D layout of the place, to build assistive navigation systems. We propose a novel line-based image global descriptor that encloses the structure of the scene observed. This descriptor is designed with omnidirectional imagery in mind, where observed lines are longer than in conventional images. Our experiments show that the proposed descriptor can be used for indoor scene recognition comparing its results to state-of-the-art global descriptors. Besides, we demonstrate additional advantages of particular interest for wearable vision systems: higher robustness to rotation, compactness, and easier integration with other scene understanding steps.

Improving Robot Mobility by Combining Downward-Looking and Frontal Cameras

This paper presents a novel attempt to combine a downward-looking camera and a forward-looking camera for terrain classification in the field of off-road mobile robots. The first camera is employed to identify the terrain beneath the robot. This information is then used to improve the classification of the forthcoming terrain acquired from the frontal camera. This research also shows the usefulness of the Gist descriptor for terrain classification purposes. Physical experiments conducted in different terrains (quasi-planar terrains) and different lighting conditions, confirm the satisfactory performance of this approach in comparison with a simple color-based classifier based only on frontal images. Our proposal substantially reduces the misclassification rate of the color-based classifier (∼10% versus ∼20%).

3D Spatial Layout Propagation in a Video Sequence

Intelligent autonomous systems need detailed models of their environment to achieve sophisticated tasks. Vision sensors provide rich information and are broadly used to obtain these models, particularly, indoor scene understanding has been widely studied. A common initial step to solve this problem is the estimation of the \(3\)D layout of the scene. This work addresses the problem of scene layout propagation along a video sequence. We use a Particle Filter framework to propagate the scene layout obtained using a state-of-the-art technique on the initial frame and propose how to generate, evaluate and sample new layout hypotheses on each frame. Our intuition is that we can obtain better layout estimation at each frame through propagation than running separately at each image. The experimental validation shows promising results for the presented approach.

Building an Enhanced Vocabulary of the Robot Environment with a Ceiling Pointing Camera.

Mobile robots are of great help for automatic monitoring tasks in different environments. One of the first tasks that needs to be addressed when creating these kinds of robotic systems is modeling the robot environment. This work proposes a pipeline to build an enhanced visual model of a robot environment indoors. Vision based recognition approaches frequently use quantized feature spaces, commonly known as Bag of Words (BoW) or vocabulary representations. A drawback using standard BoW approaches is that semantic information is not considered as a criteria to create the visual words. To solve this challenging task, this paper studies how to leverage the standard vocabulary construction process to obtain a more meaningful visual vocabulary of the robot work environment using image sequences. We take advantage of spatio-temporal constraints and prior knowledge about the position of the camera. The key contribution of our work is the definition of a new pipeline to create a model of the environment. This pipeline incorporates (1) tracking information to the process of vocabulary construction and (2) geometric cues to the appearance descriptors. Motivated by long term robotic applications, such as the aforementioned monitoring tasks, we focus on a configuration where the robot camera points to the ceiling, which captures more stable regions of the environment. The experimental validation shows how our vocabulary models the environment in more detail than standard vocabulary approaches, without loss of recognition performance. We show different robotic tasks that could benefit of the use of our visual vocabulary approach, such as place recognition or object discovery. For this validation, we use our publicly available data-set.

Line-based global descriptor for omnidirectional vision

Scene understanding is a widely studied problem in computer vision. Many works approach this problem in indoor environments assuming constraints about the scene, such as the typical Manhattan World assumption. The goal of this work is to design and evaluate a global descriptor for indoor panoramic images that encloses information about the 3D structure. This descriptor is based on the detection of representative lines of the scene, which encode the scene structure. Our work focuses on omnidirectional imagery, where observed lines are longer than in conventional images and the whole scene is captured in a single image. Experiments using two public datasets analyze the performance of the descriptor for scene categorization. We also analyze the influence of different parameters and show sample results for a navigation assistance application.

3D Layout Propagation to Improve Object Recognition in Egocentric Videos

Intelligent systems need complex and detailed models of their environment to achieve more sophisticated tasks, such as assistance to the user. Vision sensors provide rich information and are broadly used to obtain these models, for example, indoor scene modeling from monocular images has been widely studied. A common initial step in those settings is the estimation of the \(3\)D layout of the scene. While most of the previous approaches obtain the scene layout from a single image, this work presents a novel approach to estimate the initial layout and addresses the problem of how to propagate it on a video. We propose to use a particle filter framework for this propagation process and describe how to generate and sample new layout hypotheses for the scene on each of the following frames. We present different ways to evaluate and rank these hypotheses. The experimental validation is run on two recent and publicly available datasets and shows promising results on the estimation of a basic \(3\)D layout. Our experiments demonstrate how this layout information can be used to improve detection tasks useful for a human user, in particular sign detection, by easily rejecting false positives.