Ichiroh Kanaya, Japan | ISBN: | Submitted: Jun 30, 1998

Three-Dimensional Modeling of Relic Piece for Virtual Relic Restoration

We propose a new approach to simulate restoration of relics using the Virtual Reality technology.

We first designed a new method to build three-dimensional modeling of shards (broken relic pieces) using a laser range finder. Laser rangefinder is a kind of relatively accurate and efficient three-dimensional digitizer than other kinds of three-dimensional digitizer like MRI equipment in terms of archaeological use. However, we cannot measure whole shape (front and back surface) of a shard at once using a typical plane-scan type laser rangefinder.

Thus, we first developed a view integration method for data of relic shards obtained by a laser rangefinder. By our approach, a single relic shard is scanned only twice, the front surface and the back surface, by a plane-scan type laser rangefinder. And then the two shape data of front and back surface are integrated under the physical constraint of the shard shape.

The integration of the shape data is done as follows. We searched the ground-contact points on the back surface data of the shard. In this search process, we simulated physical constraint of the shard to determine which points on the back surface of the shard touch to the ground. Next, we calculate the ground-plane, the plane that contains the ground-contact points, of the back surface data. We perform three-dimensional rotation of the data of the back surface so that the ground-plane agrees the real global ground plane (the floor).

Finally we compute gravity centers and inertia of the silhouette images of both surface data. Two-dimensional Affine transformation is performed to the front surface data so that the gravity center and the inertia of the front surface data agree the gravity center and the inertia of the back surface data.

Texture mapping is also done on the virtually reconstructed relic shard (virtual shard). The texture image is taken by a digital video camera and mapped onto the virtual relic. Two-dimensional linear projection is performed to the original texture image so that the texture is correctly mapped to the virtual relic.

We examined these reconstruction methods for several real relics. The experiment has shown that the reconstructed virtual relics were enough accurate and realistic for archaeological use.

We then build virtual relic restoration environment using computer graphics technology. The stereo monitor and three-dimensional positioning equipment are used for the virtual restoration.

The experiment to examine the accuracy of virtual restoration against the real restoration environment has shown that enough accuracy and efficiency of our approach for archaeological study.

Furthermore, once the real shards are digitized and reconstructed on a computer, we will have the following advantage against the direct use of real shards: 1) Archaeologists can redo restoration. 2) They can place any shards at any position (because the virtual environment no longer has gravity). 3) They can observe shards from any directions. 4) They can store virtual shards on database, which can be distributed via computer network like Internet.

[more information]

[other authors]
Yuko Kanemoto, Qian Chen and Kunihiro Chihara

[keywords]
Information Archaeology, Restoration of relics, 3D modeling, View integration