ITC/USA 2014 Conference Proceedings / The Fiftieth Annual International Telemetering Conference and Technical Exhibition / October 20-23, 2014 / Town and Country Resort & Convention Center, San Diego, CA / Flight tests and flight safety are tightly coupled. For a given experimental test flight, the main objective to be achieved is related to the test bed and crew integrity (i.e. Bring back the test bed in one piece). Furthermore, the aircraft operational safety (i.e. continued airworthiness) relies into the accuracy of gathered information, provided by the airborne Flight Test Instrumentation (FTI) system. Typically, the FTI system includes: a data acquisition system; a transmission system; a data recorder; multiple sensors and auxiliary equipment such as Global Navigation Satellite System (GNSS) Receiver. In a specific Flight Test Campaign (FTC), such as external stores separation FTC, the associated risk is high because the non-linear and unpredicted aerodynamics couplings between the released store and the aircraft structure (e.g. wings) could lead to a catastrophic collision condition. Typically, the setup for such FTC is a big technical challenge for the instrumentation group because the determination of six degrees of freedom (6Dof) trajectory of the release store becomes mandatory. Such information could be by the means of an optical tracking system that requires the integration of two or more hi-speed hi-resolution video cameras into FTI. Unfortunately, the FTI and the Real-Time Telemetry Link (RTL) limited bandwidth imposes several restrictions for the development and implementation of a real-time application to be used at the Ground Telemetry System (GTS) for the verification of the separation safety and the validation of the test point. This paper, presents the implementation of computer vision techniques and Camshift algorithm as an approach to tracking individual track targets on the store surface which can be perform onboard and in real time or near real time. The development of the algorithm is presented as well several experimental test results that use videos frames gathered from previous FTC executed by Instituto de Pesquisas e Ensaios em Voo (Flight Test and Research Institute - IPEV). Preliminary results present satisfactory performance.

In this dissertation, novel variational optical-flow and active-contour methods are investigated to address challenging problems in computer vision and shape-based imaging. Starting from traditional applications of these methods in computer vision, such as object segmentation, tracking, and detection, this research subsequently applies similar active contour techniques to the realm of shape-based imaging, which is an image reconstruction technique estimating object shapes directly from physical wave measurements. In particular, the first and second part of this thesis deal with the following two physically inspired computer vision applications. Optical Flow for Vision-Based Flame Detection: Fire motion is estimated using optimal mass transport optical flow, whose motion model is inspired by the physical law of mass conservation, a governing equation for fire dynamics. The estimated motion fields are used to first detect candidate regions characterized by high motion activity, which are then tracked over time using active contours. To classify candidate regions, a neural net is trained on a set of novel motion features, which are extracted from optical flow fields of candidate regions. Coupled Photo-Geometric Object Features: Active contour models for segmentation in thermal videos are presented, which generalize the well-known Mumford-Shah functional. The diffusive nature of heat processes in thermal imagery motivates the use of Mumford-Shah-type smooth approximations for the image radiance. Mumford-Shah's isotropic smoothness constraint is generalized to anisotropic diffusion in this dissertation, where the image gradient is decomposed into components parallel and perpendicular to level set curves describing the object's boundary contour. In a limiting case, this anisotropic Mumford-Shah segmentation energy yields a one-dimensional ``photo-geometric'' representation of an object which is invariant to translation, rotation and scale. These properties allow the photo-geometric object representation to be efficiently used as a radiance feature; a recognition-segmentation active contour energy, whose shape and radiance follow a training model obtained by principal component analysis of a training set's shape and radiance features, is finally applied to tracking problems in thermal imagery. The third part of this thesis investigates a physics-driven active contour approach for shape-based imaging. Adjoint Active Contours for Shape-Based Imaging: The goal of this research is to estimate both location and shape of buried objects from surface measurements of waves scattered from the object. These objects' shapes are described by active contours: A misfit energy quantifying the discrepancy between measured and simulated wave amplitudes is minimized with respect to object shape using the adjoint state method. The minimizing active contour evolution requires numerical forward scattering solutions, which are obtained by way of the method of fundamental solutions, a meshfree collocation method. In combination with active contours being implemented as level sets, one obtains a completely meshfree algorithm; a considerable advantage over previous work in this field. With future applications in medical and geophysical imaging in mind, the method is formulated for acoustic and elastodynamic wave processes in the frequency domain.

published_or_final_version / abstract / Electrical and Electronic Engineering / Master / Master of Philosophy

published_or_final_version / abstract / Computer Science / Master / Master of Philosophy

published_or_final_version / abstract / toc / Electrical and Electronic Engineering / Master / Master of Philosophy

published_or_final_version / abstract / toc / Electrical and Electronic Engineering / Master / Master of Philosophy

published_or_final_version / Electrical and Electronic Engineering / Master / Master of Philosophy

published_or_final_version / Computer Science and Information Systems / Master / Master of Philosophy

Ο πρωταρχικός στόχος αυτής της εργασίας είναι η υλοποίηση μιας βιομηχανικής διάταξης σε εργαστηριακή κλίμακα στην οποία να συνεργάζονται ένα στερεοσκοπικό σύστημα και ένας ρομποτικός βραχίονας. Πιο συγκεκριμένα, η διαδικασία χωρίζεται σε δύο μέρη, την αναγνώριση στο χώρο των επιθυμητών αντικειμένων και την οδήγηση βάση αυτής του ρομποτικού βραχίονα. Για το πρώτο μέρος έγινε χρήση της βιβλιοθήκης OpenCV σε γλώσσα C++ και ως στερεοσκοπικό υλικό χρησιμοποιήθηκαν δύο παράλληλα διατεταγμένες κάμερες τύπου webcam. Η διαδικασία που ακολουθήθηκε για την αναγνώριση θέσης χωρίστηκε σε αρκετά βήματα. Αρχικά δημιουργήθηκε ένα πρόγραμμα το οποίο αποθηκεύει καρέ από τις δύο κάμερες στα οποία απεικονίζεται ένα μοτίβο βαθμονόμησης. Στη συνέχεια, αυτές οι εικόνες εισάγονται στον κώδικα βαθμονόμησης με στόχο να υπολογιστούν οι εγγενείς και εξωγενείς παράμετροι των καμερών. Έπειτα, με τη χρήση των παραμέτρων αυτών και τη θεωρία της επιπολικής γεωμετρίας μπορεί να γίνει η αναγνώριση θέσης ενός αντικειμένου στο χώρο. Τέλος, χρησιμοποιείται ένας αλγόριθμος εντοπισμού του κέντρου ενός αντικειμένου στην οθόνη με βάση το χρώμα έτσι ώστε να καθοριστεί για ποιο αντικείμενο ο αλγόριθμος θα υπολογίσει τη θέση. Στο δεύτερο μέρος χρησιμοποιήθηκε ο ρομποτικός βραχίονας Katana s400 6M90G της εταιρείας Neuronics, ο οποίος προγραμματίστηκε σε γλώσσα C++, σε περιβάλλον Visual Studio 2008. Αρχικά, βρέθηκαν οι γωνίες Euler της αρπάγης, για διαφορετικές προσεγγίσεις του προσανατολισμού της. Με τον συνδυασμό των συντεταγμένων του αντικειμένου, που βρίσκονται από το στερεοσκοπικό σύστημα καθοδηγείται το Katana ώστε να το πιάσει. Τα πειράματα που διεξαχθήκαν περιλάμβαναν την αρπαγή στάσιμων αντικειμένων με γενικό και οριζόντιο προσανατολισμό εργαλείου. Τέλος, πραγματοποιήθηκαν πειράματα με αντικείμενα εν κινήσει, τυχαίας θέσης με γενικό, κάθετο και οριζόντιο προσανατολισμό αρπάγης. / The primary objective of this thesis is the implementation of an industrial system at laboratory scale in which a stereo system collaborates with a robotic arm. More specifically, the process is divided into two parts, the recognition of the desired objects and the manipulation of the robotic arm. In the first part were used the OpenCV library, in C ++ language and two parallel web-cameras. The procedure for the recognition of the objects, was divided into several steps. Initially a number of images were inserted in the calibration algorithm in order to estimate the intrinsic and extrinsic camera parameters. Then, using these parameters and the epipolar geometry it was possible to recognize the position of an object in 3D space. Finally, an algorithm is used to locate the center of an object on the screen by color in order to determine for what object the algorithm will calculate the position. In the second part was used the robotic arm Katana s400 6M90G by Neuronics, programmed in C ++, in Visual Studio 2008. Initially, the Euler angles of the gripper for different orientations were found. Given the coordinates of the objects, provided by the stereo system, the Katana arm was guided to grasp them. The experiments that were conducted included the grasping of stationary objects in a general and horizontal orientation of the Katana tool. Finally, experiments were performed with objects in motion and random position in general, vertical and horizontal orientation of the gripper.

Στόχος της παρούσας διπλωματικής εργασίας είναι η μελέτη και η ανάπτυξη ενός ρομποτικού οχήματος το οποίο θα είναι ικανό να πλοηγηθεί αυτόνομα στο χώρο αναγνωρίζοντας και αποφεύγοντας τα πιθανά εμπόδια στο περιβάλλον που θα κινηθεί. / The aim of this thesis is to study and develop a robotic vehicle that will be able to navigate autonomously in space by identifying and avoiding potential barriers in the environment that it will move.