This study introduces a thermal defect detection technique for power equipment based on a spectral transformation model. First, the spectral residual transform model is constructed according to the redundancy of the natural background and significance of the thermal defect target in infrared images of power equipment. Then, the infrared image of the power equipment is transformed by spectral residuals to remove redundant image information of the natural background target, and a result map with significant information is generated. The experimental results show that compared with other traditional thermal defect detection methods, the proposed method has significant advantages in terms of recognition accuracy and efficiency and meets the application requirements of thermal fault detection of power equipment.