Indoor unmanned aerial vehicle (UAV) inspection in substations can effectively reduce the intensity of manual inspection operations. Due to high flight accuracy requirements and limited carrying capacity, relying solely on UAV-mounted cameras and inertial measurement unit (IMU) data fusion to determine pose fails to meet precision requirements. Therefore, a multi-vision-inertial navigation fusion framework based on the ubiquitous IoT with existing fixed cameras in the substation is proposed. The images of UAV-mounted cameras for indoor lighting conditions are enhanced and combined with IMU data to obtain preliminary UAV position data. In addition, by deploying quick response (QR) codes on UAVs, the improved perspective?n?point (PnP) algorithm is applied to optimize UAV pose data. After the flight is completed, the cumulative error of IMU in the UAV nest is verified. Experimental results have shown that the deployment and maintenance workload of this method is small, and the flight accuracy is significantly improved compared to relying solely on cameras and IMU data fusion algorithms. It can meet the needs of UAV inspection operations in substations.