Scheme for Improving the Spatial Resolution of Thermal Images for Objects with Different Emissivity

Abstract

In this study, we propose a method for acquiring a spatial high-resolution thermal distribution from a low-resolution infrared thermal camera while correcting the temperature based on emissivity. Thermal images depend on the color and material of the measurement target. Therefore, temperature cannot directly be measured accurately by infrared thermal camera. This is a serious problem for measuring the temperature of an electrical terminal carrying a large current in electrical equipment monitoring. In this study, we improved the temperature profile of targets by example-based super-resolution, in other words, by creating a dictionary comprising pairs of low-resolution and high-resolution thermal images corrected by emissivity. We experimentally proved that temperature profiles were improved, that is, the temperature profiles were close to the measurements obtained by thermocouples and high-resolution infrared thermal camera.