Vision-Based Tracking, Odometry and Control for UAV Autonomy

El principal objetivo de este trabajo es proporcionar una solución en tiempo real basada en visión estéreo o monocular precisa y robusta para que un vehículo aéreo no tripulado (UAV) sea autónomo en varios tipos de aplicaciones UAV, especialmente en entornos abarrotados sin señal GPS. Este trabajo principalmente consiste en tres temas de investigación de UAV basados en técnicas de visión por computador: (I) visual tracking, proporciona soluciones efectivas para localizar visualmente objetos de interés estáticos o en movimiento durante el tiempo que dura el vuelo del UAV mediante una aproximación adaptativa online y una estrategia de múltiple resolución, de este modo superamos los problemas generados por las diferentes situaciones desafiantes, tales como cambios significativos de aspecto, iluminación del entorno variante, fondo del tracking embarullado, oclusión parcial o total de objetos, variaciones rápidas de posición y vibraciones mecánicas a bordo. La solución ha sido utilizada en aterrizajes autónomos, inspección de plataformas mar adentro o tracking de aviones en pleno vuelo para su detección y evasión; (II) odometría visual: proporciona una solución eficiente al UAV para estimar la posición con 6 grados de libertad (6D) usando únicamente la entrada de una cámara estéreo a bordo del UAV. Un método Semi-Global Blocking Matching (SGBM) eficiente basado en una estrategia grueso-a-fino ha sido implementada para una rápida y profunda estimación del plano. Además, la solución toma provecho eficazmente de la información 2D y 3D para estimar la posición 6D, resolviendo de esta manera la limitación de un punto de referencia fijo en la cámara estéreo. Una robusta aproximación volumétrica de mapping basada en el framework Octomap ha sido utilizada para reconstruir entornos cerrados y al aire libre bastante abarrotados en 3D con memoria y errores correlacionados espacialmente o temporalmente; (III) visual control, ofrece soluciones de control prácticas para la navegación de un UAV usando Fuzzy Logic Controller (FLC) con la estimación visual. Y el framework de Cross-Entropy Optimization (CEO) ha sido usado para optimizar el factor de escala y la función de pertenencia en FLC. Todas las soluciones basadas en visión en este trabajo han sido probadas en test reales. Y los conjuntos de datos de imágenes reales grabados en estos test o disponibles para la comunidad pública han sido utilizados para evaluar el rendimiento de estas soluciones basadas en visión con ground truth. Además, las soluciones de visión presentadas han sido comparadas con algoritmos de visión del estado del arte. Los test reales y los resultados de evaluación muestran que las soluciones basadas en visión proporcionadas han obtenido rendimientos en tiempo real precisos y robustos, o han alcanzado un mejor rendimiento que aquellos algoritmos del estado del arte. La estimación basada en visión ha ganado un rol muy importante en controlar un UAV típico para alcanzar autonomía en aplicaciones UAV. ABSTRACT The main objective of this dissertation is providing real-time accurate robust monocular or stereo vision-based solution for Unmanned Aerial Vehicle (UAV) to achieve the autonomy in various types of UAV applications, especially in GPS-denied dynamic cluttered environments. This dissertation mainly consists of three UAV research topics based on computer vision technique: (I) visual tracking, it supplys effective solutions to visually locate interesting static or moving object over time during UAV flight with on-line adaptivity approach and multiple-resolution strategy, thereby overcoming the problems generated by the different challenging situations, such as significant appearance change, variant surrounding illumination, cluttered tracking background, partial or full object occlusion, rapid pose variation and onboard mechanical vibration. The solutions have been utilized in autonomous landing, offshore floating platform inspection and midair aircraft tracking for sense-and-avoid; (II) visual odometry: it provides the efficient solution for UAV to estimate the 6 Degree-of-freedom (6D) pose using only the input of stereo camera onboard UAV. An efficient Semi-Global Blocking Matching (SGBM) method based on a coarse-to-fine strategy has been implemented for fast depth map estimation. In addition, the solution effectively takes advantage of both 2D and 3D information to estimate the 6D pose, thereby solving the limitation of a fixed small baseline in the stereo camera. A robust volumetric occupancy mapping approach based on the Octomap framework has been utilized to reconstruct indoor and outdoor large-scale cluttered environments in 3D with less temporally or spatially correlated measurement errors and memory; (III) visual control, it offers practical control solutions to navigate UAV using Fuzzy Logic Controller (FLC) with the visual estimation. And the Cross-Entropy Optimization (CEO) framework has been used to optimize the scaling factor and the membership function in FLC. All the vision-based solutions in this dissertation have been tested in real tests. And the real image datasets recorded from these tests or available from public community have been utilized to evaluate the performance of these vision-based solutions with ground truth. Additionally, the presented vision solutions have compared with the state-of-art visual algorithms. Real tests and evaluation results show that the provided vision-based solutions have obtained real-time accurate robust performances, or gained better performance than those state-of-art visual algorithms. The vision-based estimation has played a critically important role for controlling a typical UAV to achieve autonomy in the UAV application.

Image recognition: Visual grouping, recognition, and learning

Vision extracts useful information from images. Reconstructing the three-dimensional structure of our environment and recognizing the objects that populate it are among the most important functions of our visual system. Computer vision researchers study the computational principles of vision and aim at designing algorithms that reproduce these functions. Vision is difficult: the same scene may give rise to very different images depending on illumination and viewpoint. Typically, an astronomical number of hypotheses exist that in principle have to be analyzed to infer a correct scene description. Moreover, image information might be extracted at different levels of spatial and logical resolution dependent on the image processing task. Knowledge of the world allows the visual system to limit the amount of ambiguity and to greatly simplify visual computations. We discuss how simple properties of the world are captured by the Gestalt rules of grouping, how the visual system may learn and organize models of objects for recognition, and how one may control the complexity of the description that the visual system computes.

Object-related activity revealed by functional magnetic resonance imaging in human occipital cortex.

The stages of integration leading from local feature analysis to object recognition were explored in human visual cortex by using the technique of functional magnetic resonance imaging. Here we report evidence for object-related activation. Such activation was located at the lateral-posterior aspect of the occipital lobe, just abutting the posterior aspect of the motion-sensitive area MT/V5, in a region termed the lateral occipital complex (LO). LO showed preferential activation to images of objects, compared to a wide range of texture patterns. This activation was not caused by a global difference in the Fourier spatial frequency content of objects versus texture images, since object images produced enhanced LO activation compared to textures matched in power spectra but randomized in phase. The preferential activation to objects also could not be explained by different patterns of eye movements: similar levels of activation were observed when subjects fixated on the objects and when they scanned the objects with their eyes. Additional manipulations such as spatial frequency filtering and a 4-fold change in visual size did not affect LO activation. These results suggest that the enhanced responses to objects were not a manifestation of low-level visual processing. A striking demonstration that activity in LO is uniquely correlated to object detectability was produced by the "Lincoln" illusion, in which blurring of objects digitized into large blocks paradoxically increases their recognizability. Such blurring led to significant enhancement of LO activation. Despite the preferential activation to objects, LO did not seem to be involved in the final, "semantic," stages of the recognition process. Thus, objects varying widely in their recognizability (e.g., famous faces, common objects, and unfamiliar three-dimensional abstract sculptures) activated it to a similar degree. These results are thus evidence for an intermediate link in the chain of processing stages leading to object recognition in human visual cortex.

Large scale SLAM with visual features

Comunicación presentada en el X Workshop of Physical Agents, Cáceres, 10-11 septiembre 2009.

Human-computer interaction for optical music recognition tasks

The need to digitise music scores has led to the development of Optical Music Recognition (OMR) tools. Unfortunately, the performance of these systems is still far from providing acceptable results. This situation forces the user to be involved in the process due to the need of correcting the mistakes made during recognition. However, this correction is performed over the output of the system, so these interventions are not exploited to improve the performance of the recognition. This work sets the scenario in which human and machine interact to accurately complete the OMR task with the least possible effort for the user.

3D visual sensing of human hand for remote operation of a robotic hand

New low cost sensors and open free libraries for 3D image processing are making important advances in robot vision applications possible, such as three-dimensional object recognition, semantic mapping, navigation and localization of robots, human detection and/or gesture recognition for human-machine interaction. In this paper, a novel method for recognizing and tracking the fingers of a human hand is presented. This method is based on point clouds from range images captured by a RGBD sensor. It works in real time and it does not require visual marks, camera calibration or previous knowledge of the environment. Moreover, it works successfully even when multiple objects appear in the scene or when the ambient light is changed. Furthermore, this method was designed to develop a human interface to control domestic or industrial devices, remotely. In this paper, the method was tested by operating a robotic hand. Firstly, the human hand was recognized and the fingers were detected. Secondly, the movement of the fingers was analysed and mapped to be imitated by a robotic hand.

A Performance Evaluation of Surface Normals-Based Descriptors for Recognition of Objects using CAD-Models

This paper describes a study and analysis of surface normal-base descriptors for 3D object recognition. Specifically, we evaluate the behaviour of descriptors in the recognition process using virtual models of objects created from CAD software. Later, we test them in real scenes using synthetic objects created with a 3D printer from the virtual models. In both cases, the same virtual models are used on the matching process to find similarity. The difference between both experiments is in the type of views used in the tests. Our analysis evaluates three subjects: the effectiveness of 3D descriptors depending on the viewpoint of camera, the geometry complexity of the model and the runtime used to do the recognition process and the success rate to recognize a view of object among the models saved in the database.

Visual perception for the 3D recognition of geometric pieces in robotic manipulation

During grasping and intelligent robotic manipulation tasks, the camera position relative to the scene changes dramatically because the robot is moving to adapt its path and correctly grasp objects. This is because the camera is mounted at the robot effector. For this reason, in this type of environment, a visual recognition system must be implemented to recognize and “automatically and autonomously” obtain the positions of objects in the scene. Furthermore, in industrial environments, all objects that are manipulated by robots are made of the same material and cannot be differentiated by features such as texture or color. In this work, first, a study and analysis of 3D recognition descriptors has been completed for application in these environments. Second, a visual recognition system designed from specific distributed client-server architecture has been proposed to be applied in the recognition process of industrial objects without these appearance features. Our system has been implemented to overcome problems of recognition when the objects can only be recognized by geometric shape and the simplicity of shapes could create ambiguity. Finally, some real tests are performed and illustrated to verify the satisfactory performance of the proposed system.

Genetic algorithm optimization of adaptive multi-scale GLCM features

We introduce a new second-order method of texture analysis called Adaptive Multi-Scale Grey Level Co-occurrence Matrix (AMSGLCM), based on the well-known Grey Level Co-occurrence Matrix (GLCM) method. The method deviates significantly from GLCM in that features are extracted, not via a fixed 2D weighting function of co-occurrence matrix elements, but by a variable summation of matrix elements in 3D localized neighborhoods. We subsequently present a new methodology for extracting optimized, highly discriminant features from these localized areas using adaptive Gaussian weighting functions. Genetic Algorithm (GA) optimization is used to produce a set of features whose classification worth is evaluated by discriminatory power and feature correlation considerations. We critically appraised the performance of our method and GLCM in pairwise classification of images from visually similar texture classes, captured from Markov Random Field (MRF) synthesized, natural, and biological origins. In these cross-validated classification trials, our method demonstrated significant benefits over GLCM, including increased feature discriminatory power, automatic feature adaptability, and significantly improved classification performance.

The correspondence framework for 3D surface matching algorithms

Beyond the inherent technical challenges, current research into the three dimensional surface correspondence problem is hampered by a lack of uniform terminology, an abundance of application specific algorithms, and the absence of a consistent model for comparing existing approaches and developing new ones. This paper addresses these challenges by presenting a framework for analysing, comparing, developing, and implementing surface correspondence algorithms. The framework uses five distinct stages to establish correspondence between surfaces. It is general, encompassing a wide variety of existing techniques, and flexible, facilitating the synthesis of new correspondence algorithms. This paper presents a review of existing surface correspondence algorithms, and shows how they fit into the correspondence framework. It also shows how the framework can be used to analyse and compare existing algorithms and develop new algorithms using the framework's modular structure. Six algorithms, four existing and two new, are implemented using the framework. Each implemented algorithm is used to match a number of surface pairs. Results demonstrate that the correspondence framework implementations are faithful implementations of existing algorithms, and that powerful new surface correspondence algorithms can be created. (C) 2004 Elsevier Inc. All rights reserved.

A First Order Predicate Logic Formulation of the 3D Reconstruction Problem and its Solution Space

This paper defines the 3D reconstruction problem as the process of reconstructing a 3D scene from numerous 2D visual images of that scene. It is well known that this problem is ill-posed, and numerous constraints and assumptions are used in 3D reconstruction algorithms in order to reduce the solution space. Unfortunately, most constraints only work in a certain range of situations and often constraints are built into the most fundamental methods (e.g. Area Based Matching assumes that all the pixels in the window belong to the same object). This paper presents a novel formulation of the 3D reconstruction problem, using a voxel framework and first order logic equations, which does not contain any additional constraints or assumptions. Solving this formulation for a set of input images gives all the possible solutions for that set, rather than picking a solution that is deemed most likely. Using this formulation, this paper studies the problem of uniqueness in 3D reconstruction and how the solution space changes for different configurations of input images. It is found that it is not possible to guarantee a unique solution, no matter how many images are taken of the scene, their orientation or even how much color variation is in the scene itself. Results of using the formulation to reconstruct a few small voxel spaces are also presented. They show that the number of solutions is extremely large for even very small voxel spaces (5 x 5 voxel space gives 10 to 10(7) solutions). This shows the need for constraints to reduce the solution space to a reasonable size. Finally, it is noted that because of the discrete nature of the formulation, the solution space size can be easily calculated, making the formulation a useful tool to numerically evaluate the usefulness of any constraints that are added.

A global vision system for a robot soccer team

This paper describes the real time global vision system for the robot soccer team the RoboRoos. It has a highly optimised pipeline that includes thresholding, segmenting, colour normalising, object recognition and perspective and lens correction. It has a fast ‘paint’ colour calibration system that can calibrate in any face of the YUV or HSI cube. It also autonomously selects both an appropriate camera gain and colour gains robot regions across the field to achieve colour uniformity. Camera geometry calibration is performed automatically from selection of keypoints on the field. The system acheives a position accuracy of better than 15mm over a 4m × 5.5m field, and orientation accuracy to within 1°. It processes 614 × 480 pixels at 60Hz on a 2.0GHz Pentium 4 microprocessor.

Low-cost real-time gesture recognition

A major impediment to developing real-time computer vision systems has been the computational power and level of skill required to process video streams in real-time. This has meant that many researchers have either analysed video streams off-line or used expensive dedicated hardware acceleration techniques. Recent software and hardware developments have greatly eased the development burden of realtime image analysis leading to the development of portable systems using cheap PC hardware and software exploiting the Multimedia Extension (MMX) instruction set of the Intel Pentium chip. This paper describes the implementation of a computationally efficient computer vision system for recognizing hand gestures using efficient coding and MMX-acceleration to achieve real-time performance on low cost hardware.

Mach edges: a critical test of the nonlinear 3rd derivative model for edge-detection

Feature detection is a crucial stage of visual processing. In previous feature-marking experiments we found that peaks in the 3rd derivative of the luminance profile can signify edges where there are no 1st derivative peaks nor 2nd derivative zero-crossings (Wallis and George 'Mach edges' (the edges of Mach bands) were nicely predicted by a new nonlinear model based on 3rd derivative filtering. As a critical test of the model, we now use a new class of stimuli, formed by adding a linear luminance ramp to the blurred triangle waves used previously. The ramp has no effect on the second or higher derivatives, but the nonlinear model predicts a shift from seeing two edges to seeing only one edge as the added ramp gradient increases. In experiment 1, subjects judged whether one or two edges were visible on each trial. In experiment 2, subjects used a cursor to mark perceived edges and bars. The position and polarity of the marked edges were close to model predictions. Both experiments produced the predicted shift from two to one Mach edge, but the shift was less complete than predicted. We conclude that the model is a useful predictor of edge perception, but needs some modification.