Digitization method for mining key elements of cultural relics

Abstract

The invention provides a digitization method for mining key elements of cultural relics, and aims to support the protection work of the cultural relics through a high-precision point cloud file generation technology. According to the method, triangular laser scanning, structured light scanning, photogrammetry and other technologies are combined, the high-dimensional-precision overall shape and contour of the cultural relic can be rapidly obtained, and meanwhile point cloud data with high texture detail precision and high color precision are obtained. The multi-technology fusion method not only improves the data processing efficiency, but also ensures the high precision of the finally generated point cloud model in size, detail and color. In addition, the method further comprises a grid model generation step based on the point cloud model, and convenience is provided for follow-up utilization and development of the cultural relic digital model. The method has advancement and practicability in technology, and is of great significance in improving the efficiency and quality of cultural relic protection and promoting inheritance and development of cultural relics.

Description

Technical Field

[0001] The present invention belongs to the field of cultural relics protection, and specifically relates to technologies such as 3D scanning, image processing, and point cloud data processing. Background Art

[0002] With the continuous progress of technology, three-dimensional scanning technology has become an important tool in the field of cultural relics protection. Traditional methods of cultural relics protection often rely on manual measurement and manual repair, which are not only time-consuming and laborious, but also prone to causing secondary damage to cultural relics. Therefore, a new technology that can efficiently and accurately obtain the three-dimensional information of cultural relics is particularly important.

[0003] Although existing three-dimensional scanning technologies have solved the above problems to a certain extent, there are still some deficiencies. For example, many scanning devices cannot adapt to the complex and changeable surfaces of cultural relics, resulting in insufficiently fine point cloud data and unable to meet the requirements of high-precision restoration. In addition, existing software is inefficient in processing large amounts of point cloud data and cannot achieve rapid modeling and restoration.

[0004] To solve these problems, the present invention provides a digital method for mining key elements of cultural relics, which can generate fine point cloud files and provide strong technical support for cultural relics protection. While ensuring high precision, this method greatly improves the efficiency of data processing. The core of this method lies in the combination of multiple scanning technologies, including triangular laser scanning, structured light scanning, photogrammetry and other technologies. This multi-technology fusion method can give full play to the advantages of various technologies and make up for the deficiencies of individual technologies. Through low-precision scanning, point cloud data of the overall shape and contour of the cultural relic with high dimensional accuracy can be quickly obtained; through structured light high-precision scanning, point cloud data with high texture detail accuracy can be obtained; through photogrammetry, point cloud data with high color accuracy can be obtained. The combination of these technologies enables the method of the present invention to generate a point cloud model with both high dimensional accuracy, high detail fineness, and high color accuracy. At the same time, this method also provides steps for generating a mesh model based on the point cloud model for the subsequent utilization and development of the cultural relic digital model.

[0005] In summary, this method not only has technological advancement and practicality, but also has important application value in the field of cultural relics protection. By implementing the present invention, the efficiency and quality of cultural relics protection can be effectively improved, making a positive contribution to the inheritance and development of cultural relics. Summary of the Invention

[0006] The present invention is a digital method for mining key elements of cultural relics based on multi-source three-dimensional scanning technology. This method is specifically implemented by adopting the following technical solutions:

[0007] 1. Point cloud data acquisition

[0008] (1) Low-resolution overall scanning

[0009] Use structured light three-dimensional scanning or triangulation laser scanning technology to scan the whole cultural relic object. The scanning resolution is 0.5 - 1.0 mm. In this step, a rough point cloud model of the whole cultural relic object is obtained. Due to the limitations of structured light three-dimensional scanning and triangulation laser scanning technologies, this point cloud model does not contain color information but has high geometric accuracy.

[0010] (2) High-resolution local scanning

[0011] Use structured light three-dimensional scanning or triangulation laser scanning technology to scan the local complex texture of the cultural relic object. The scanning resolution is 0.05 - 0.2 mm, and then use the spherical rotation algorithm to generate a mesh. In this step, a local texture point cloud model of the cultural relic object is obtained. This point cloud model does not contain color information but has extremely high geometric accuracy.

[0012] (3) Color scanning

[0013] Use photogrammetry technology to take photos around the scanning object. It is necessary to ensure that the scene light is uniform and soft during this process. In this step, image data of each angle of the cultural relic is obtained. Next, perform color correction, image mask production, image alignment, and dense point cloud generation on the image data of each angle of the cultural relic in sequence. This step can capture the spatial geometric information and color information on the surface of the cultural relic object simultaneously, obtaining an overall color point cloud model. Due to the limitations of photogrammetry technology, the geometric accuracy of this model is low, but it contains high-precision color information.

[0014] 2. Point cloud data processing

[0015] (1) Add texture to the rough overall point cloud model

[0016] First, perform point cloud alignment. Use a feature-based alignment method to extract features such as corner points, edges, and curvature extreme points of the rough overall point cloud model and the local texture point cloud model for point cloud alignment. Then use the iterative closest point algorithm to further adjust and optimize the alignment result to obtain a more refined alignment result. Then perform point cloud fusion, covering the data at the same position in the rough overall point cloud model with the point cloud data in the local texture point cloud model to obtain an overall texture point cloud model.

[0017] (2) Add color information to the overall texture point cloud model

[0018] Align the overall color point cloud model and the overall texture point cloud model using the same method as in (1). Then, use the nearest neighbor interpolation method to assign the color information of each point in the overall color point cloud model to the corresponding points in the overall texture point cloud model, obtaining a high-precision point cloud model of the cultural relic. This model contains both the fine texture feature information and color information of the cultural relic, and can accurately digitally record the three-dimensional features of the cultural relic.

[0019] 3. Mesh Data Generation

[0020] Based on the high-precision point cloud model, generate a high-precision mesh model. First, perform regional segmentation to divide the point cloud model into a textureless region and a textured region. Then, perform mesh generation. Use the Poisson reconstruction algorithm to generate a mesh for the textureless region. The mesh generated by this algorithm is relatively smooth. Use the ball rolling algorithm to generate a mesh for the textured region. This process can accurately retain the texture information and has a relatively small computer load. Finally, perform mesh fusion. Since the meshes generated based on the same point cloud file are already in the same coordinate system, they can be directly merged, and algorithms such as Laplacian Smoothing or Taubin's Smoothing are used for edge smoothing to obtain a fine mesh model containing color information. Description of the Drawings

[0021] Figure 1 For the scanning time, processing time, and obtained data graphs of different scanning technologies;

[0022] Figure 2 For the comparison graph of color differences of different scanning devices Detailed Implementation Manner

[0023] The present invention will be further described below in conjunction with embodiments:

[0024] Embodiment 1

[0025] This case performs scanning work on three cultural relics: a Qing Dynasty flower window, a Qing Dynasty lattice door, and a late Qing Dynasty lattice door.

[0026] 1. Point Cloud Data Acquisition

[0027] (1) Low-resolution overall scanning and high-resolution local scanning

[0028] These two steps were carried out twice using the equipment of two technical routes to compare their effects. The first scheme was carried out using the TrackScan-P (TSP) equipment based on triangular laser scanning technology. The resolution was set to 0.5 mm for overall scanning, and then 0.2 mm was used for local scanning of the decorative parts with complex textures. The second scheme was carried out using the Artec Eva and Artec Space Spider (AE&ASS) equipment based on structured light three-dimensional scanning technology. Each of the two equipment had its own focus. AE focused on large-size and low-precision scanning work, so AE was used for overall scanning. ASS focused on small-size and high-precision scanning work for local scanning. In addition, the Agisoft Metashape software (AM) based on photogrammetry technology was used with a camera to carry out this work for subsequent performance comparison.

[0029] (2) Color scanning

[0030] This step was carried out using AM with a camera. In addition, AE&ASS was also used for this work for subsequent performance comparison.

[0031] 2. Point cloud data processing and mesh data generation

[0032] These steps were completed using software such as Geomagic Wrap and Artec studio, which can perform operations such as point cloud denoising, point cloud alignment, point cloud merging, and mesh generation.

[0033] 3. Performance evaluation

[0034] (1) Geometric accuracy evaluation

[0035] This evaluation compared the data volumes of the point cloud models obtained by various methods to characterize the geometric accuracy, as Figure 1 shown.

[0036] Figure 1 are the scanning time, processing time, and obtained data graphs for different scanning technologies. It can be seen that the point cloud data obtained by triangular laser scanning and structured light three-dimensional scanning are almost the same in geometric accuracy and far superior to photogrammetry technology. This evaluation confirms the rationality of using structured light three-dimensional scanning or triangular laser scanning technology for low-resolution overall scanning and high-resolution local scanning in this method.

[0037] (2) Color accuracy evaluation

[0038] This evaluation uses the X-rite SP60 color difference meter as a reference to evaluate the color scanning accuracy of three devices, namely AM, AE, and ASS. Since the TSR based on triangular laser scanning technology cannot scan color information, it is not within the scope of evaluation. The scanning accuracy is characterized by the color difference (ΔE), which is obtained from the LAB data derived from the color difference meter. The calculation formula is:

[0039] ΔE = (ΔL 2 + ΔA 2 + ΔB 2 ) 1 / 2

[0040] The evaluation results are as Figure 2 shown in Figure 2 the comparison chart of color differences of different scanning devices. It can be seen that the color difference value of AM based on photogrammetry is significantly smaller than that of the other two devices based on structured light scanning technology. This evaluation verifies the rationality of using photogrammetry technology for color scanning.

Claims

1. A digital method for mining key elements of cultural relics, comprising the following steps: 1.1 Data collection 1.1.1 Low-resolution overall scanning This step is based on structured light three-dimensional scanning or triangulation laser scanning technology, with a scanning resolution of 0.5 - 1.0 mm, to obtain a rough overall point cloud model of the cultural relic object. 1.1.2 High-resolution local scanning This step is based on structured light three-dimensional scanning or triangulation laser scanning technology, with a scanning resolution of 0.05 - 0.2 mm, to obtain a local texture point cloud model of the cultural relic object. 1.1.3 Color scanning This step is based on photogrammetry technology. Photos are taken around the scanning object. Next, color correction, image masking, image alignment, and dense point cloud generation are performed on the captured image data in sequence to obtain an overall color point cloud model. 1.2 Point cloud data processing 1.2.1 Adding texture to the overall rough point cloud model Align the overall rough point cloud model and the local texture point cloud model based on the feature-based alignment method and the iterative closest point algorithm. Then, use the point cloud data in the local texture point cloud model to overwrite the data at the same position in the overall rough point cloud model to obtain an overall texture point cloud model. 1.2.2 Adding color information to the overall texture point cloud model Align the overall color point cloud model and the overall texture point cloud model using the same method as in 1.2.

1. Then, use the nearest neighbor interpolation method to assign the color information of each point in the overall color point cloud model to the corresponding point in the overall texture point cloud model to obtain a high-precision point cloud model of the cultural relic object containing color information. 1.3 Mesh data generation Based on the high-precision point cloud model, a high-precision mesh model is generated. First, regional segmentation is performed to divide the point cloud model into a textureless region and a textured region. Then, mesh generation is carried out. The Poisson reconstruction algorithm is used for mesh generation in the textureless region, and the ball rolling algorithm is used for mesh generation in the textured region. Finally, mesh fusion and fusion edge smoothing processing are performed.

Images