The reconstruction of the complete view of a virtual environment has been paid substantial attention by the researchers. It has long been recognized that the large-scale object reconstruction is an important enabling technology for realistic, large-scale visualization of the world. The goal of this research is synthesizing the whole views of an object from the sampled viewpoint images of a scene. Especially, a full view image of the large- scale object is reconstructed. The whole view of a large-scale object is impossible to take the whole scene at once without using sufficient distance between camera and object.
The system of the whole view synthesizing is developed base on the image registration approach. Image registration technique is very useful for computing the transformation and constructing a panoramic image from the sampled images of a scene. The multiple images are grabbed by a camera system from different camera viewpoints. Feature extraction algorithm is developed for extracting the feature points from each image. Histogram equalization and lighting adjusting are performed to improve the accuracy of feature extraction. The corresponding feature point pairs among the successive image pairs are determined from the relation of the camera positions. The perspective projected geometry method is applied for detecting the relation between the camera and image coordinate system from the image pairs. In this approach, it is possible to produce a lot of intermediate views by using a few cameras. Then it will be provided to reduce the cost and number of camera in security system by using the synthesizing the arbitrary view of images.
The application form and entrance exam structure can be downloaded ...
The application form and entrance exam structure can be downloaded ...
Result for Ph.D (IT) and Ph.D (Research Only) ...