Detection of Faulty Power Apparatuses Using Distance Information and Temperature Information

When an apparatus which should be diagnosed is near a neighboring heat source, it is difficult to distinguish the apparatus from the other heat source from only a thermal image, because there is no distance information. Therefore, we added distance information obtained with a stereo camera to temperature information obtained with an infrared camera, and developed a method of detecting a faulty apparatus on distribution poles using the infrared and stereo cameras. This method has four subcomponents: (1) generation of distance information with a stereo camera, (2) projection of distance information space onto temperature information space and integration of distance information and temperature information, (3) extraction of apparatuses on a thermal image, which is a visualization of temperature information, using distance information, and (4) detection of a faulty apparatus based on the local temperature gradient. We verified this method experimentally. The experimental result of using the new method indicated that by integrating distance information and temperature information, the effect of neighboring heat sources was removed and the faulty apparatus could be correctly detected.