Dimitrios Chrysostomou

Fault diagnosis of photovoltaic modules through image processing and Canny edge detection on field thermographic measurements

Today, conventional condition monitoring of installed, operating photovoltaic (PV) modules is mainly based on electrical measurements and performance evaluation. However, such practices exhibit restricted fault-detection ability. This study proposes the use of standard thermal image processing and the Canny edge detection operator as diagnostic tools for module-related faults that lead to hot-spot heating effects. The intended techniques were applied on thermal images of defective PV modules, from several field infrared thermographic measurements conducted during this study. The whole approach provided promising results with the detection of hot-spot formations that were diagnosed to specific defective cells in each inspected module. These evolving hot spots lead to abnormally low performance of the PV modules, a fact that is also validated by the manufacturers standard electrical tests.

A fuzzy multi-sensor architecture for indoor navigation

This paper presents an indoor navigation system based on sensor data from first responder wearable modules. The proposed system integrates data from an inertial sensor, a digital camera and a radio frequency identification device using a sophisticated fuzzy algorithm. To improve the navigation accuracy, different types of first responder activities and operational conditions were examined and classified according to extracted qualitative attributes. The vertical acceleration data, which indicates the periodic vibration during gait cycle, is used to evaluate the accuracy of the inertial based navigation subsystem. The amount of strong feature correspondences assess the quality of the three-dimensional scene knowledge from digital camera feedback. Finally, the qualitative attribute, in order to evaluate the efficiency of the radio frequency identification subsystem, is the degree of probability of each location estimate. Fuzzy if-then rules are then applied to these three attributes in order to carry out the fusion task. Simulation results based on the proposed architecture have shown better navigation effectiveness and lower positioning error compared with the used stand alone navigation systems.

A biologically inspired scale-space for illumination invariant feature detection

This paper presents a new illumination invariant operator, combining the nonlinear characteristics of biological center-surround cells with the classic difference of Gaussians operator. It specifically targets the underexposed image regions, exhibiting increased sensitivity to low contrast, while not affecting performance in the correctly exposed ones. The proposed operator can be used to create a scale-space, which in turn can be a part of a SIFT-based detector module. The main advantage of this illumination invariant scale-space is that, using just one global threshold, keypoints can be detected in both dark and bright image regions. In order to evaluate the degree of illumination invariance that the proposed, as well as other, existing, operators exhibit, a new benchmark dataset is introduced. It features a greater variety of imaging conditions, compared to existing databases, containing real scenes under various degrees and combinations of uniform and non-uniform illumination. Experimental results show that the proposed detector extracts a greater number of features, with a high level of repeatability, compared to other approaches, for both uniform and non-uniform illumination. This, along with its simple implementation, renders the proposed feature detector particularly appropriate for outdoor vision systems, working in environments under uncontrolled illumination conditions.

Improving the robustness in feature detection by local contrast enhancement

This paper presents a new feature detector, with improved local contrast performance. The proposed method is based on an improved non-linear version of the classic Difference of Gaussians, which exhibits increased sensitivity to low contrast. Additionally, it does not employ computationally expensive or memory demanding routines. In order to evaluate the degree of illumination invariance that the proposed, as well as, other existing detectors exhibit, a new benchmark image database has been created. It features a greater variety of imaging conditions, compared to existing databases, containing real scenes under various degrees and combinations of uniform and non-uniform illumination. Experimental results show that the proposed detector extracts greater number of features, with high level of repeatability, compared to other existing ones. These results are evident for both uniform and non-uniform illumination, evincing a favorable usage of the proposed feature detector by robotic platforms working in outdoor working environments.

Methods and techniques for intelligent navigation and manipulation for bomb disposal and rescue operations

Handing a teleoperated robotic mechanism demands special skills and involves particular problems. Especially in cases of robots dealing with rescue operations or bomb disposal. In such cases any lost in communications might arise unpredictable results. Also either a bomb or a survivor need attentional handling. In this paper we describe automatic methods and techniques developed on a multifunctional teleoperated robot. These intend to assist both the robot and the human operator in accomplishing their mission towards rescue or bomb disposal.

Optimum multi-camera arrangement using a bee colony algorithm

This paper presents a new method for designing multi-camera arrangements with aim to maximize coverage with the minimum number of cameras. More specifically, the presented problem has three different components, namely (a) to maximize the coverage subject to a given number of cameras (b) to optimize the camera topology given fixed locations and (c) to minimize the cost of the arrangement, while the least required percentage of coverage is provided. In order to solve these problems, a bee colony algorithm is utilized as an optimization technique that is able to determine the minimum number of cameras needed to cover the given space completely while taking into consideration the minimum possible cost for the proposed arrangement as well. The algorithm employs several camera placement constraints referring to geometrical, optical as well as reconstructive limitations and delivers promising preliminary results.

A bio-inspired multi-camera system for dynamic crowd analysis

A new way to maximize the coverage of an examined floor plan in a dynamic fashion.Bio-inspired bee and spider agents for evolving crowd analysis.Utilize simulated pedestrian data to observe the density of crowd areas.Three new challenging floor plans are now examined.Three algorithms are compared along the proposed method for their performance. Analysis of crowd density has emerged nowadays as a hot topic issue related to the crowd safety and comfort and directly depended on the design and the operation of the crowded places under study. Usually multiple camera networks are employed to cover, monitor and improve the safety of people in large multifunctional crowded buildings. On the other hand, the art gallery problem is a computational geometry approach to a classical real-world visibility challenge. In a nutshell, it concerns the minimization of the free moving guards required to observe the entire gallery. In this paper we attempt to approach this problem from a novel perspective. To begin with, the number of guards are replaced by multiple cameras whose number should be minimized. At the same time, the observability of the camera network in the available space should be dynamically maximized, so as to observe the evolving density of the crowded areas adequately. In order to achieve this objective a twofold bio-inspired method is described and implemented, based on the emergent computation of swarms to come up with solutions in complex mathematical problems. More specifically, the observations on bumblebee colonies lead us firstly to the definition of artificial bumblebee agents used to determine the number of cameras needed to maximize the observability of a space given the safety specifications emerged from the crowd analysis. Secondly, the way the spiders wave their webs was used as a source of inspiration to determine the exact positions of the cameras in the given space by artificial spider agents. The feedback of the algorithm is then used to cover the areas with significant crowd density in a dynamic fashion. Experimental results show that the algorithm is capable of producing promising results where the areas with the maximum crowd density are continuously detected and covered in a dynamic way.

Obtaining Reliable Depth Maps for Robotic Applications from a Quad-Camera System

Autonomous navigation behaviors in robotics often require reliable depth maps. The use of vision sensors is the most popular choice in such tasks. On the other hand, accurate vision-based depth computing methods suffer from long execution times. This paper proposes a novel quad-camera based system able to calculate fast and accurately a single depth map of a scenery. The four cameras are placed on the corners of a square. Thus, three, differently oriented, stereo pairs result when considering a single reference image (namely an horizontal, a vertical and a diagonal pair). The proposed system utilizes a custom tailored, simple, rapidly executed stereo correspondence algorithm applied to each stereo pair. This way, the computational load is kept within reasonable limits. A reliability measure is used in order to validate each point of the resulting disparity maps. Finally, the three disparity maps are fused together according to their reliabilities. The maximum reliability is chosen for every pixel. The final output of the proposed system is a highly reliable depth map which can be used for higher level robotic behaviors.

Multi-view 3D scene reconstruction using ant colony optimization techniques

This paper presents a new method performing high-quality 3D object reconstruction of complex shapes derived from multiple, calibrated photographs of the same scene. The novelty of this research is found in two basic elements, namely: (i) a novel voxel dissimilarity measure, which accommodates the elimination of the lighting variations of the models and (ii) the use of an ant colony approach for further refinement of the final 3D models. The proposed reconstruction procedure employs a volumetric method based on a novel projection test for the production of a visual hull. While the presented algorithm shares certain aspects with the space carving algorithm, it is, nevertheless, first enhanced with the lightness compensating image comparison method, and then refined using ant colony optimization. The algorithm is fast, computationally simple and results in accurate representations of the input scenes. In addition, compared to previous publications, the particular nature of the proposed algorithm allows accurate 3D volumetric measurements under demanding lighting environmental conditions, due to the fact that it can cope with uneven light scenes, resulting from the characteristics of the voxel dissimilarity measure applied. Besides, the intelligent behavior of the ant colony framework provides the opportunity to formulate the process as a combinatorial optimization problem, which can then be solved by means of a colony of cooperating artificial ants, resulting in very promising results. The method is validated with several real datasets, along with qualitative comparisons with other state-of-the-art 3D reconstruction techniques, following the Middlebury benchmark.

Generalization of orientation trajectories and force-torque profiles for robotic assembly

Abstract A typical robot assembly operation involves contacts with the parts of the product to be assembled and consequently requires the knowledge of not only position and orientation trajectories but also the accompanying force-torque profiles for successful performance. To learn the execution of assembly operations even when the geometry of the product varies across task executions, the robot needs to be able to adapt its motion based on a parametric description of the current task condition, which is usually provided by geometrical properties of the parts involved in the assembly. In our previous work we showed how positional control policies can be generalized to different task conditions. In this paper we propose a complete methodology to generalize also the orientational trajectories and the accompanying force-torque profiles to compute the necessary control policy for a given condition of the assembly task. Our method is based on statistical generalization of successfully recorded executions at different task conditions, which are acquired by kinesthetic guiding. The parameters that describe the varying task conditions define queries into the recorded training data. To improve the execution of the skill after generalization, we combine the proposed approach with an adaptation method, thus enabling the refinement of the generalized assembly operation.