Visualization of landscape by VRML system

Tsuyoshi Honjo and En-Mi Lim

Graduate School of Science and Technology, Chiba University

1-33 Yayoi-cho, Inage-ku, Chiba, 263 Japan

(CorrespondenceName: Tsuyoshi Honjo Phone and fax : +81-47-308-8896 E-mail: honjo@midori.h.chiba-u.ac.jp

Address: Faculty of Horticulture, Chiba University 648 Matsudo, Matsudo-shi, Chiba-ken, Japan, 271-8510)

Source Reference

Honjo, T. and Lim, E., 2001. Visualization of landscape by VRML system. Landscape and Urban Planning, 55, 175-183.

Abstract

VRML (Virtual Reality Modeling Language) is a high performance language for 3-D visualization on the WWW (World Wide Web). Three-dimensional information can be easily transferred through the Internet by this technology. In this study, we made a landscape visualization system that enables virtual experience in a planned landscape by using VRML and the applicability of the system to the landscape design was shown. To perform real time rendering of landscape, a tree was expressed by using 2 textured planes instead of thousands of polygons. Trees were placed automatically on a textured terrain based on the plant database by the system developed in this study. With the system, we made models of real gardens based on measured data and walk-through simulations in the gardens were tested. The system showed good performance and it also indicated the potential of VRML systems. The information on the virtual landscape can be placed on WWW. This method can be utilized both for the design of and for the public discussion on landscape planning.

Keywords: Computer Graphics; Landscape design; Visualization; Virtual Reality; VRML

1. Introduction

In landscape planning, simulation of the landscape is a powerful tool for public understanding and for selection of alternative planning scenerios. In the simulation, reality of the simulated image is very important and recent progress of computer graphics enables very precise simulation of the landscape.

　 In the computer graphics of landscape simulation, quality of plants plays an important role. Many researchers have studied realistic plant modeling (Honda, 1971; Aono and Kunii, 1984; Bloomenthal, 1985; Oppenheimer, 1986; Prusinkiewicz et al., 1988; De Reffye et al., 1988; Greene, 1989; Viennot et al., 1989). By using techniques of plant modeling, very realistic images have been made for landscape planning and estimation (Honjo et al., 1992; Saito et al., 1993; Morimoto, 1993; Honjo and Takeuchi, 1995). In these cases of landscape simulation, still images were mainly utilized because the numbers of polygons that comprise simulated scenery are very large. Walk-through animation or real time rendering of the landscape were difficult. Animations were made as sequences of the still images, by changing the eye points and viewed points.

In this study, we made a system by using Virtual Reality Modeling Language (VRML) and the applicability of the system for landscape visualization was evaluated. The VRML system enables real-time virtual experience of walk-through simulation in a planned landscape. VRML is also a high performance language for three dimensional (3-D) visualization on the World Wide Web (WWW) and 3-D information can be easily transferred through the Internet (Honjo et al., 1997; Honjo and Takeuchi, 1998).

Application of virtual reality to landscape planning will make possible the precise recognition of the plan but needs real-time rendering of the landscape according to the moving viewpoint. Formerly, to realize the real-time rendering, virtual reality systems consisted of input and output devices and a high-performance computer was necessary. On the other hand, to make virtual reality environment on the Internet with VRML, it is possible to make the environment on a personal computer economically.

In landscape planning, the feed back process from user to planner shown in Fig. 1 is important. The VRML system is suitable for this feed back process because walk-through simulation on the Internet contributes in the user’s understanding of the planning (Honjo et al., 1999; Lim et al., 2000).

Fig. 1 Process of landscape planning using visualization by VRML Fig. 2 Process of landscape simulation using VRML

2. Methods

2.1 VRML systems

　 VRML is a programming language and library for 3-D computer graphics and has many functions. The first version was made in 1994 as VRML 1.0 and the second version, VRML 2.0, which has more dynamic functions, was made in 1996. In this study VRML 2.0 was used to make the system.

To make the rendering very fast, VRML supports only simple rendering techniques such as shading, setting objects, projection, and texture mapping but does not support complicated rendering such as ray tracing. Programming by VRML is easier than that by a graphic library like OpenGL.

　 The user who downloads the program from a server can use a program made by VRML. A 3-D image made by VRML is rendered on the local computer of the user. To use the VRML, a browser that supports VRML is necessary. In this study, Cosmo Player (Silicon Graphics Inc.) was used as VRML browser with Internet Explorer (Microsoft Inc.).

For VRML programming, the VRML browser and an Internet browser are necessary. Programming and landscape planning is possible on a stand-alone computer. Cosmo Player and other VRML browsers can be used as freeware and the developing environment can be built very economically.

2.2VRML and visualization of landscape

Basic functions for the visualization of landscape are the modeling and setting of terrain, plants, and architecture. The procedures of the visualization are shown in Fig 2.

To validate the landscape planning, many perspective images should be made and discussed. By using the VRML system, once the program of the landscape is made, changing the viewpoint is easy and walk-through simulation is possible.

Some CAD software has a function that transfers CAD data to VRML format. However, the software does have the function which is necessary for the landscape simulation. For example, making a VRML file from collected research data on trees by changing the expression according to the growth of the trees. In this study, we developed a system that was optimized for landscape visualization and automatically made the VRML program.

2.3 Modeling of terrain

In VRML2.0 when there are elevation data (DEM) on a grid, terrain is easily visualized by using a node (command used in VRML) called ElevationGrid.

When the elevation data are not on a grid, grid data for elevation is calculated by interpolating the original elevation data. In this study, the interpolation program was made in Visual Basic (Microsoft Inc.).

By mapping a texture on the terrain data, the quality of the reality of the terrain is improved. The texture mapping is also used with the ElevationGrid node.

2.4 Modeling of plants and architecture

To make a fast rendering of plants, the texture of plants which are in a transparent GIF format are mapped on a plane and two planes are crossed to show the plants as shown in Fig. 3. This method is very effective for the fast rendering of plants.

For a landscape simulation that consists of plants at various growth stages, texture images of plants of each growth stages are required.

We used computer graphics images of plants made by AMAP (Atelier de Modelisation de Architecture de Plants), which is a system (developed by CIRAD, Center Internationale Recherche Agricultural Development) that produces high precision 3-D plant shapes. AMAP is one of the outputs of the study of plants by DE Reffye et al. (1988). AMAP generates very botanically realistic 3-D computer graphics. By AMAP, 3-D computer graphics images of several growth stages can be easily made and are used as a texture.

3-D Plants made by computer graphics by AMAP or other techniques consist of polygons. Branch, twigs, flowers and leaves are all described by sets of polygons. The number of the polygons varies between thousands to millions. Such a polygon model is suitable for photo realistic expression of the plants but needs large amount of time for the rendering and additionally, walk-through simulation is difficult in VRML. Therefore, we used only 2-D textures in this study and a 2-D plant image database was developed. Examples of plant images in the database are shown in Fig. 4.

　 To make architecture such as buildings, simple objects such as cubes were used to reduce the number of polygons and texture mapping was used for adding reality in this study.

2.5 Converting plant investigation data to VRML

In the investigation of plant data in forests and gardens, species, location of plants or density of plants, height, width and diameter of trunk (usually at breast height) were recorded.

We converted the investigated data to VRML format. For the conversion, a program was developed in Visual Basic (Microsoft Inc.).

Fig. 3 Example of trees images made by 2 textured planes. Fig. 4. Examples of plant images in the database.

3. Results

3.1 Visualization of a small garden

By using the developed landscape planning system, a VRML program was made from collected data and the performance of the system was evaluated. VRML images were compared with photographs to gauge the performance of the program.

A garden in the Faculty of Horticulture in Chiba University was simulated. A map of the garden is shown in Fig. 5. There were 184 plants in the garden and the species and their heights were measured. The data were converted to a VRML format.

In the garden of VRML, the number of the polygons was approximately 400 and the amount of data was approximately 400 KB (Honjo et al., 1999). Walk-through simulation was very smooth with this data (with a computer of Cerelon 300MHz).

Quality of the VRML images was generally precise compared to photographs. In Fig.6, a sequence of images of a walk-through simulation is shown. Similarity to reality and recognition of objects were very good because of the smooth walk-through simulation.

Fig. 5 Map of the garden in Chiba University. Fig. 6 Images of walk-through animation the garden in Chiba University.

3.2 Visualization historical Japanese garden

Koishikawa Korakuen Garden, one of the most famous Japanese gardens, was simulated. The data of the garden were mainly taken from a research report on the environment of the garden in 1985 (Tokyo Metropolitan Government and Korakuen Stadium, 1986). The first procedure for the landscape simulation was the creation of terrain data. A map of 1:400 made in the environmental research for the garden was entered in a computer by a scanner and locations of control points of were measured from the image. 2051 control points were read from the image and used to make the terrain in VRML (Fig. 7).

The number of the plants in the garden was 3956 consisting 40 families, 67 genera and 98 species. For all the plants, species, height, diameter at breast height (about at 120 cm) and location were investigated and recorded. In this study, we used 2980 plants with 91 species whose heights were more than 3m and girths at breast height were more than 30 cm. The locations of these plants are shown in Fig. 8. The data for plants were prepared from an image database. In this case, data in summer were used.

In Fig. 9 the landscape of the garden was simulated and it is compared with photographs.

The size of the VRML program was 4.34 MB (1.61 MB for plant data, 2.05 MB for terrain data and 0.68 MB for texture data of plants). Although the amount of data was large, viewpoints can be changed in a few seconds and walk-through simulation was possible (Lim et al., 2000).

With the system developed in this study, we proved that landscape design with some thousands of plants was possible and that the system can be used as a practical and low cost landscape design system.

3.3 Communication on the WWW

The VRML programs made in this study were placed on a WWW server. Apache was used as the WWW server program. The landscape data developed in this study can be opened and be accessed by anyone. In the planning of landscape, using WWW to transfer information to the public is considered as a very important and practical method.

Fig. 7 Measured control points terrain (2051 points). Fig. 8 Measured location of the plants (2980 plants, 91 species).

Fig. 9 Comparison of photos and VRML images of the Koishikawa Korakuen Garden.

Fig. 10 Prediction of landscape change in the future.

4. Discussions

4.1 Amount of data

If we consider the access to the VRML file, the data should be within the size of smooth transfer on the network. For users, a compact VRML program with a small number of polygons and texture data also enables a smooth walk-through simulation.

In the near future, faster network speed and a high performance computer will enable the transfer of large amounts of data.

4.2 Image database

A database of 2-D plant images with various species and growth stages was developed. Mostly the images generated by AMAP were used and the quality of the images was very high. We made a plant image database that includes more than 1000 images in this study.

4.3 Prediction of future landscape

In this study, the system that visualizes landscape in 3-D can also be used for the future prediction of the landscape. With the system, we can simulate not only the present landscape but also the future landscape. In Fig. 10, future sceneries of the garden, which is shown in Fig. 6, are predicted. Some plants are changed from the original garden and a lapse of 15 years and its effect reflect the landscape of the garden.

4.4 Expansion of the function of the system

The system made in this study has a function to transfer research data to a VRML program. This system can be broadly used for design tools for the planning of landscapes, that include plants, terrain and architecture. The system can also be used for a 3-D presentation tool of the landscapes.

We made only a simple graphic user interface. As a further expansion of the function of the system, we are developing graphical user interface. By adding graphical user interface, the system can be more easily used as a CAD system for 3-D design of landscape visualization. It also enables interactive design and allows the designer to view the future landscape under various design scenarios as shown in Fig. 10.

4.5 Applied areas of the system

A number of additional applied areas of the VRML system can be considered. These areas include assessing user preference for designs, finding a solution from alternative designs and quantifying urban green spaces more precisely. By connecting the plant growth and administrative costs, it will be possible to predict the future operating costs of the designed landscape. It is also a good educational tool for landscape architects.

5. Conclusions

In this study, VRML was used for developing a landscape design system and its availability and possibilities were evaluated. A database of plant images with various species and growth stages was also developed. With the system, real gardens were simulated based on measured data and it was shown that 3-D landscape design was possible. Real-time rendering of the landscape and walk-through simulation in the garden was also possible. The 3-D information can be easily transferred through the Internet by this technology.

The system showed good performance and it also indicated the potential of building a Virtual Reality system by VRML with very low cost. The information on the virtual landscape can be opened on the WWW. This method should be utilized in public discussion about landscape planning.

Examples of virtual gardens shown in this paper can be accessed at the following URL.

http://leo.h.chiba-u.ac.jp

References

Aono, M. and Kunii, T., 1984. Botanical Tree Image Generation. IEEE Computer Graphics & Applications, 4, 10-34.

Bloomenthal, J., 1985. Modeling the Mighty Maple. Computer Graphics, 19, 305-311.

Greene, N., 1989. Voxel Space Automata:Modeling with Stochastic Growth Processes in Voxel Space, Computer Graphics, 23, 175-184

Honda, H., 1971. Description of the Form of Trees by the Parameters of the Tree-like Body:Effects of the Branching Angle and the Branch Length on Shape of the Tree-like Body. J. Theor. Biol., 31, 331-338.

Honjo, T., Saito, K. and Kumagai Y., 1992. A Study of the Application of Plant Modeling and Its Visualization on Landscape Prediction. Journal of the Japanese Institute of Landscape Architecture, 55(5), 301-306. (In Japanese with English summary)

Honjo, T. and Takeuchi, S., 1995. Visualization of Local Environment. The Bulletin of School of High-Tech. Tokai Univ., Vol. 5, 177-181. (In Japanese with English summary)

Honjo, T., Hoshi, T., Takeuchi, S. and Fuse, J., 1997. Visualization System of Environmental Information on Internet. Journal of the Japanese Institute of Landscape Architecture, 60(5), 611-614. (In Japanese with English summary)

Honjo, T. and Takeuchi, S., 1998. Method for Three Dimensional Terrain Visualization on Internet. Journal of the Japanese Institute of Landscape Architecture, 61(5), 605-608. (In Japanese with English summary)

Honjo, T., Lim, E. and Maruta Y., 1999. Landscape Visualization by VRML. Journal of the Japanese Institute of Landscape Architecture, 62(5), 665-668. (In Japanese with English summary)

Lim, E., Honjo, T. and Maruta Y., 2000. The Study of Landscape Visualization by VRML and Plant Modeling Database. Journal of the Japanese Institute of Landscape Architecture, 63(5), 799-802. (In Japanese with English summary)

Morimoto, Y., 1993. Computer Visualization of the Vegetation of Katsura Imperial Garden by a Plant Modeling and Visualization System. Journal of the Japanese Institute of Landscape Architecture, 57(5), 113-120. (In Japanese with English summary)

Oppenheimer, P. E., 1986. Real Time Design and Animation of Fractal Plants and Trees. Computer Graphics, 20, 55-64.

Prusinkiewicz, P., Lindenmayer, A. and Hanan, J., 1988. Developmental Models of Herbaceous Plants for Computer Imagery Purposes. Computer Graphics, 22, 146-160.

De Reffye, P., Edelin, C., Francon, J., Jaeger, M. and Puech, C., 1988. Plant Models Faithful to Botanical Structure and Development. Computer Graphics, 22, 151-158.

Saito, K., Honjo, T., Kumagai Y., Ishida, Y., Lecoustre, R. and De Reffye, P.，1993. Photo-realistic Forest Landscape Simulation –Application of GIS and plant modeling technique-. Nicongraph Paper Contest, Vol. 9, 226-236. (In Japanese with English summary)

Tokyo Metropolitan Government and Korakuen Stadium, 1986: Report on environmental and biological investigation of Koishikawa Korakuen in 1985. (In Japanese)

Viennot, X. G., Eyrolles, G., Janey, N. and Arques, D., 1989. Combinatorial Analysis of Ramified Patterns and Computer Imageryof Trees. Computer Graphics, 23, 31-40.