Grant-in-Aid for Scientific Research (B)
|Allocation Type||Single-year Grants |
|Research Institution||Nara Institute of Science and Technology |
YOKOYA Naokazu Graduate School of Information Science, Nara Institute of Science and Technology, Professor, 情報科学研究科, 教授 (10252834)
YAMAZAWA Kazumasa Graduate School of Information Science, Nara Institute of Science and Technology, Research Associate, 情報科学研究科, 助手 (40283931)
IWASA Hidehiko Graduate School of Information Science, Nara Institute of Science and Technology, Research Associate, 情報科学研究科, 助手 (50263447)
TAKEMURA Haruo Graduate School of Information Science, Nara Institute of Science and Technology, Associate Professor, 情報科学研究科, 助教授 (60263430)
ARAKI Shoichi Matsushita Electric Industrial Co., Ltd., Senior Researcher, 先端技術研究所, 主任研究員
ETOH Minoru Matsushita Electric Industrial Co., Ltd., Senior Researcher, 東京通信システム研究所, 主席研究員
栄籐 稔 松下電器産業(株), 東京通信システム研究所, 主席研究員
|Project Period (FY)
1997 – 1999
Completed (Fiscal Year 1999)
|Budget Amount *help
¥7,900,000 (Direct Cost: ¥7,900,000)
Fiscal Year 1999: ¥1,200,000 (Direct Cost: ¥1,200,000)
Fiscal Year 1998: ¥2,500,000 (Direct Cost: ¥2,500,000)
Fiscal Year 1997: ¥4,200,000 (Direct Cost: ¥4,200,000)
|Keywords||Omnidirectional Image Sensor / Omnidirectional Image / Cylindrical Panoramic Image / Telepresence / Virtual Reality / Computer Vision / Real-time Systems / Immersive Visual Environment|
We have obtained the following research results in this project.
1.Development of Omnidirectional Stereo Imaging Sensors.
We have developed two types of omnidirectional stereo imaging sensors. The first one is based on using a composite mirror of two hyperboloidal mirrors and a standard video camera. It has been concluded that this sensor suffers from low resolution of images and thus requires high-resolution video cameras for practical applications. The second from sensor is developed for overcoming the low-resolution problem in the first sensor. It consists of symmetrically aligned two components, each of which is constructed of a hexagonal pyramidal mirror and six video cameras.
2.Real-time Generation of Perspective and Panoramic Images from an Omnidirectional Image
Perspective and panoramic images can be computed from an omnidirectional image because our sensor components satisfy the single viewpoint constraint. In order to execute the geometric transformation from omnidirectional imag
es to perspective and/or cylindrical panoramic images in real time, we have developed a novel algorithm which integrates coordinate transformation by software and image warping by hardware. The proposed algorithm is used for realizing two types of telepresence systems ; one uses a head-mounted display (HMD) and the other is based on a large cylindrical screen and projectors.
3.Efficient Coding of Dynamic Omnidirectional and Panoramic Images
We have developed a method for estimating structure and motion from omnidirectional image sequences as well as a method for generating novel views from a sequence of omnidirectional images obtained from a moving omnidirectional sensor.
4.Acquisition of 3D Information from Omnidirectional Stereo Images
Omnidirectional images obtained by our sensors have vertical disparities. In order to generate standard binocular stereo images with horizontal disparities, a correlation-based stereo matching algorithm is applied to omnidirectional stereo images. This may be referred to as omnidirectional stereo. Less