Analysis of quantization error in line-based stereo matching

摘要

Quantization error, generated by the quantization process of an image, is inherent in computer vision. Because, especially in stereo vision, the quantization error in 2D digital images results in position errors in a reconstructed 3D scene, it is necessary to analyze stochastically quantization error in stereo vision. In this paper, an analysis of the probability density function (pdf) of quantization error in line-based stereo matching is presented. It is shown that the theoretical pdfs of quantization errors of the reconstructed 3D point positions in line-based stereo matching show general forms compared to those of the conventional error analysis in pixel-based stereo matching. Computer simulations are observed to support the theoretical pdfs.