Infrared thermography plays an important role in the inspection of electrical installations allowing to avoid failures and breakdowns. Condition monitoring using thermal imaging techniques can be effectively achieved by keeping a safe distance from the inspected equipment. The major advantage when using these techniques is that it is not mandatory to stop the equipment during the inspection. Moreover, and since data collection is done without contact, dangerous interventions can be avoided. Any irregular areas of the compound being checked will be reflected as abnormal areas. This work aims to visualize the anomalies existing in electric equipment during inspections of many installations by means of the A40M thermo-vision imager. The segmentation results show that the binarization accuracy of the supervised SVM classifier is better than that of the supervised ANN algorithm, and it applies to both larger and smaller faults. However, the application of artificial neural networks requires more calculation time due to training and testing for verification of the region of interest. Comparison between the thermal images obtained using both methods is discussed.