A Method for Evaluating the Effect and Accuracy of 3D Reconstruction Based on Infrared Image
CHEN Jichi1， WEI Guohua2， CHAI Juanfang3
1. A Brigade of Naval Equipment Department， Beijing 100081， China；
2. School of Information and Electronics， Beijing Institute of Technology， Beijing 100081， China；
3. Shanghai Electro-Mechanical Engineering Institute， Shanghai 201109， China
Abstract Since the current 3D reconstruction evaluation indexes are only suitable for the reconstruction of target visible images, they lack attention to the overall accuracy measurement of the model and ignore the local correctness of the model, indicating that they’re not fully applicable to the 3D reconstruction evaluation of infrared images. In this study, a composite index evaluation model was proposed, including indicators such as the key feature integrity of the reconstruction model, the accuracy of camera pose estimation, and the error benchmark distance of 3D reconstruction evaluation. The simulation results show that the proposed reconstruction model evaluation method performs well robustness under different reconstruction qualities and avoids the problem that a single evaluation index cannot characterize the reconstruction accuracy of the infrared reconstruction model effectively. In addition, this proposed model can accurately quantify the reconstruction effect of point cloud and textured models, especially when compared with the existing three-dimensional reconstruction evaluation method, it is much more comprehensive.
2023, Vol. 6 
