建筑外文翻譯---研究古建筑的三維重建方法（英文）

1、Research on Method of 3D Reconstruction of Ancient Architecture (Nanputuo Temple) Junfeng Yao Software School of Xiamen University yao0010@xmu.edu.cn Hui ZHANG Software School of Xiamen University zhanghuiyc@163.com Fei

2、 SHE Software School of Xiamen University feijm_225@sina.com Abstract In this paper, regarding the Great Majesty Hall of the NanPutuo Temple as a virtual modeling object, the research on the new technology of 3D recons

3、truction of the ancient architecture has been done, in which 3DS MAX and MultiGen Creator were applied. The conflict was solved between precision and the amount of data existing in the process of 3D modeling virtual

4、 ancient architecture. The practical project has proved that the ancient architecture model made in this way has gotten living effect in roaming system, at the mean time it satisfies the data demand of real-time rend

5、ering. Keyword: Virtual reality (VR), Ancient Architecture Protection, 3DS MAX, MultiGen Creator. 1. Introduction Virtual Reality is a new man-machine interface. It provides users (participants) with satisfactory

6、immersed experience and multi-sensory channels, and attempts to seek the best way of man-machine communication[1]. Using the virtual reality technology to realize the representation of ancient architecture is the inn

7、ovation of architectural illustrational tools, and it is one of the hot topics of the present research. The lifelike computer modeling can not only demonstrate the profound cultural details of Chinese ancient archite

8、ctures, but have practical significance to the protection of ancient architectures in form of archives, therefore it was applied more widely than the traditional physical model. The modeling fidelity and rapid real-t

9、ime data rendering are the foundations of realizing virtual reality desktop system. The project “Xiamen Nanputuo Temple virtual roaming system“ required that virtual scenes could reach a balance between the quality i

10、mage and the speed of the rendering, and prescribed strictly the degree of accuracy and detail of every architecture. It controlled the amount of data as possible as we can on the basis of fine modeling. Regarding the

11、 Great Majesty Hall of the NanPutuo Temple as virtual modeling object, using the modeling method of 3DS MAX and MultiGen Creator, the paper carried on beneficial discussion to the application of the technology in geo

12、metry modeling of ancient architectures. 2. Chinese ancient architectures’ features In this paper, the virtual modeling object of the Great Majesty Hall belong to imitation Qing-Dynasty ancient architectures. It follo

13、wed the ancient Chinese architecture major characteristics generally summarized as follows [2]: (1) The most significant characteristic of ancient Chinese architecture make full use of timber framework. Wooden pos

14、ts, beams, lintels and joists make up the framework of a house. Walls serve as the separation of rooms without bearing the weight of the whole house, which is unique to China. As a famous saying goes, 'Chinese ho

15、uses will still stand when their walls collapse.' The specialty of wood requires antisepsis methods to be adopted, thus develops into Chinese own architectural painting decoration. (2) Colored glaze roofs, windo

16、ws with exquisite applique design and beautiful flower patterns on wooden pillars reflect the high-level of the craftsmen's handicraft and their rich imagination. (3) The layout of a courtyard complex is also un

17、ique to china. The main structure is located on the central axis of a court while less-important structures are located to the left and right. The whole layout is symmetrical. 3. Comparison of 3DS MAX and MultiGen Cr

18、eator In modeling of single complex architecture, fineness and fidelity, 3DS Max has more advantages more MultiGen Creator. 3DS Max provides the polygon modeling, lofting, face modeling, NURBS face modeling, and oth

19、er convenient and effective modeling method. It has a better ability of special effect International Conference on Cyberworlds 2008978-0-7695-3381-0/08 $25.00 © 2008 IEEE DOI 10.1109/CW.2008.146 627but simultaneous

20、ly also increase the burden of scene demonstration. Using the appropriate resolution textures and materials can reduce the geometric details of the model and at the same time produce better visual effects [6]. After

21、the modeling completes, process the photos taken from the spot by high resolution digital camera and then save them as texture pictures of appropriate size. Use appropriate mapping methods and coordinates in 3DS MAX

22、to implement the mapping to every component of the architecture, such as use the UVW mapping order in modifier to modify the mapping coordinate. Simultaneously to reduce the data amounts of the textures, use the repe

23、ated mapping method as far as possible in mapping. Figure 2 shows the completed 3D model of the Great Majesty Hall. Figure 2. 3D model of the Great Majesty Hall 4.4 Importing model to Creator for simplification

24、 Importing model from 3DS MAX to Creator for model simplification needs three steps: (1) Importing model. Import the 3ds file exported from 3DS MAX to Creator. Then the texture path of the model loses. So you need to

25、 designate a new texture path. The latter path should be a relative path which is the same to the model’s path. This can increase the flexibility of model path. (2) Organizing the structure of database. The database

26、 of the 3ds files does not have the hierarchical structure; all objects are placed under the identical father node. So we should re-organize the database according to the primary and secondary of different parts of th

27、e ancient architecture model. This will not only facilitate the selection operation on model in the process of modeling, but also improve the graphic rendering speed and avoid some drawing errors in the scene render

28、ing and browsing. (3) Simplifying model. That is, under the premise of not affecting the model's appearance, to display the model with few data and eliminate the redundant data. The exported 3ds files save faces

29、in form of triangle, but the Creator save faces in form of polygon. For simplification, you can use the Combine Face tool in Creator to merge the coplanar triangles to a polygonal face. This may reduce the polygon am

30、ount largely. In addition, the application of instance can simplify the model. Often, the model of a ancient architecture includes a number of same components, such as tou- kung 、 posts, beams and so on. In the modelin

31、g process, you can use the technology of instance, copy the component to other positions. So that the database will not establish the node of the same model repeatedly. This can save the memory space greatly. Finall

32、y, some invisible faces can be deleted directly, such as the underside of the model and adjacent faces between the neighboring components. 4.5 Making LOD model LOD（Level Of Details）is a model management method savin

33、g objects by different detail . The same scene or the object in the scene may call the model of different detail level according to the need of drawing scene. The higher the LOD model's detail degree is , the fin

34、er its model is, the bigger the data amount is; the lower the detail degree is , the less simple its scene or model is, the smaller the data amount is. Generally, the model with low level of details is adopted in remo

35、te distance observation. It does not affect the third dimension; when short distance observation, the model with high level of details is adopted, and you may see a very fine model [7]. There are many methods for est

36、ablishing LOD models. In the paper, we use two, deletion method and grabbing picture method [8]. The deletion method is to delete the detail parts of the model and preserve the main parts. It is suitable for the comp

37、onents of fitment level of the architecture. But the disadvantage of the method is the difficulty to minimize the model's data amount. The grabbing picture method can simplify some faces the model needs to displa

38、y, such as the windows of the architecture, their structures are very complicated. This can be done by grabbing picture method. The grabbing picture method is to grab image using the function of capturing screen, the

39、n deal with it for texture, and finally associate the texture with a whole face to demonstrate the face of the model. The faces of the completed junior LOD model should not exceed one third of the senior LOD model. F