An Accurate Modeling Technology Based on Depth Information of Cultural Relics

With the development of three-dimensional modeling technology, the requirements of digital display for cultural relics increase. This paper analyzes the inherent problems of traditional three-dimensional reconstruction techniques, that is, will generate a large number of irrelevant feature points in the process of extracting images by using traditional RGB cameras, especially for bronze artifacts, the existence of irrelevant feature points will directly affect the matching accuracy, which will cause errors in the generation of object trajectories and affect the final reconstruction effect. A method to optimize feature point matching through depth information for target extraction is proposed, thus achieving an accurate solution for accurate reconstruction. In this paper, the depth camera Kinect is used to collect images of the cultural objects with difficulty in recognition. By aligning the depth images with the RGB images, each feature point of each RGB image has depth information so that when character matching and trajectory calculation are performed, the precise and effective feature points can be screened out, that is, the feature points completely on the target object, which can significantly improve the accuracy of the matching and reduce the number of iterations in the process of Bundle Adjustment, thus greatly shortening the period of feature extraction and improving the efficiency of modeling.