This thesis proposes two methods to reconstruct the shape of 2-dimensional specular reflectors from multiple brightness images of the reflectors.Several images of a reflector are taken as the camera and a mounted light source move along the x-axis. From the specularities present in the images, reflectance information and camera locations are extracted. This information is used for reconstruction. The first method is an optimization algorithm that determines the surface shape by minimizing an objective function. The second method employs relaxation labeling. Reflector points are declared labels while the camera positions are considered objects. Object-label associations are investigated and the most consistent among these are retained to give the reflector shape.
【 预 览 】
附件列表
Files
Size
Format
View
Shape Recovery from Brightness Images of 2-Dimensional Specular Reflectors