Abstract:
Thermal cameras have widespread applications in smart grid layout optimization, equipment monitoring, and fault diagnosis
1. While these cameras visually represent temperature changes on device surfaces, they fail to provide a comprehensive view of the three-dimensional structure and heating state of the devices. In this project, some devices are equipped with both visible light and infrared cameras, but the original system only captures single infrared image for temperature monitoring. Therefore, this paper proposes a microservices-based infrared point cloud digital twin system for the fusion of single infrared image with point cloud. Firstly, an infrared image is captured by the infrared camera, the image and pose information are transmitted to infrared microservice through file microservice. Next, infrared microservice utilizes point cloud data and pose information to generate a point cloud image through rasterization method and match it's features with the infrared image. Then, the camera pose is recalculated using the PnP algorithm, and a point cloud information map is generated using the corrected pose. The results are saved as files and new version information is generated, notifying the client for updates and display through the file microservice. Through testing with real-world infrared images and point cloud data, this paper validates that the proposed method accurately restores temperature information from the two-dimensional infrared image.