Research on Overheating Detection of Cable Joints Based on Infrared Thermal Imaging and Deep Residual Network

To achieve accurate identification of thermal faults in cable joints, this study designs an overheating detection method for cable joints based on the application of infrared thermal imaging and deep residual networks. First, an infrared thermal imager is selected to capture thermal images of cable joints, and the sampled images are subjected to filtering processing to eliminate noise interference. Then, parameters of the deep residual network are designed to extract multi-level texture information from the images. Finally, through image coloring processing, intuitive detection of cable joint overheating is realized. Experimental results show that the designed method can accurately identify the overheating areas of cable joints, increase the Kappa value of detection results, and achieve the expected effect.