A high-precision thermal imaging approach based on Bayesian inference for EMC diagnosis

Electro Magnetic Compatibility (EMC) is sensitive to some of unapparent factors such as current leakage, component heating and temperature fluctuation etc. In recent years, thermal imaging techniques with high precision has been investigated for EMC diagnosis due to these unapparent factors. This paper presents a Bayesian inference based thermal imaging approach so as to calibrate temperature distortion on the surface of electronic device. The proposed method improves the thermal radiation model by studying raw data property and updating the model parameters such as measure distance, surrounding humidity, ambient temperature and surface emissivity. Thermal imaging results on ABB controllers and Schneider switches etc. validate the advantages of our proposals in the sense of high-precision, far-field detection and wide-range inspection for EMC enhancement.