立体图像视差线性调整算法

摘要

为使立体图像能在特定显示设备上舒适显示，本文提出一种调整立体图像的方法。首先利用立体图像融像区知识，计算不同显示设备舒适显示的水平视差范围；利用匹配的SIFT（ scale⁃invariant features）特征点来估计立体图像的视差范围，建立起立体图像原有视差到舒适显示视差的线性映射；最后利用奇异值分解估计该映射的变换矩阵，通过变换矩阵计算出调整后的图像，消除立体图像垂直视差，并将立体图像的水平视差调整到舒适观看的范围。实验结果表明，本文提出的方法可以解决不同显示设备上立体图像视差调整的问题，得到更好的显示效果。%A 3D image calibration method is proposed to provide a better performance on particular display device. Firstly, the knowledge of stereo image fusion area is adopted to calculate the comfortable horizontal disparity range for different display device. Secondly, the Scale⁃Invariant Features ( SIFT) algorithm is applied to estimate the disparity range in a 3D image, and then a linear mapping between the original disparity and the comfortable disparity for the 3D image are obtained. Finally, the transformational matrix for the linear mapping and disparity correction for the 3D image are got. The experimental results indicate that the proposed algorithm has a good performance, which is adaptive for various disparity problems.