This paper presents a fast and low cost methodology for thermal / visible stereo vision system calibration with high reliability for non-controlled external environments. The proposed methodology uses a non-powered calibration rig to map the coordinates, field of view and magnification distortions from visible and thermal vision and a self-calibration analysis procedure to estimate metrological uncertainty of imaging correspondence. Tests were conducted to determine the effectiveness of the developed calibration rig. 

O Problema do Caixeiro Viajante é tema de pesquisa devido a sua complexidade. Várias técnicas são constantemente pesquisadas para obter soluções aproximadas, mas eficientes. Neste trabalho são comparadas as técnicas de algoritmos genéticos, redes neurais auto-incrementáveis de Kohonen e um algoritmo heurístico proposto na literatura. Os resultados obtidos mostraram que os algoritmos genéticos apresentaram os melhores índices de acerto.

Thermography is a technique widely used for inspection of electrical equipment's operating conditions. However, its operation, predominantly manual, hampers more reliable diagnosis of component's abnormality indications. This paper presents a straightforward hybrid stereo vision configuration for autonomous monitoring systems that uses images visible and infrared spectrum to identify, characterize and extract regions of interest for increasing metrological quality of thermal measurements. Experimental results show that it is a viable and approach in asset management systems.