Main Article Content

Abstract

The majority of computational stereo methods require that the input images are rectified to make the challenges of stereo-vision more tractable. The rectification process, introduced to brings the rows of each image into alignment by warping the image captured by each viewpoint. Although rectification is often accomplished through manual calibration, automatic alignment using sparse feature-based methods is commonly performed if the stereo-head is not reachable or if frequent re-alignment is required

Article Details