Adopting deep learning and image recognition technology, it realizes intelligent, batch and automatic defect identification and diagnosis of infrared spectra collected by airborne and handheld equipment. The introduction of intelligent diagnostic technology enhances the depth and breadth of the application of image feature method, similar comparative method and file analysis method in the infrared diagnosis of equipment, especially solving the widely recognized challenge of “voltage-induced type" defect research and judgment in the industry. Deepening the supervisory level of infrared technology will bring very important engineering and practical value to improve the quality and efficiency of operation and inspection work.