Fast Stereo Matching Based on Gradient-assisted Zero-mean Normalized Cross Correlation

Zero-mean Normalized Cross Correlation (ZNCC) is commonly applied for stereo matching. However, the ZNCC-based matching could be extremely time-consuming due to its pixel-by-pixel computation. In this paper, a fast stereo matching algorithm based on gradient-assisted ZNCC is proposed. The image gradient is utilized to filter out the ineligible pixels, which reduces the number of the pixels involved in the matching cost calculation. Moreover, the gradient is incorporated into the ZNCC function to enhance the matching robustness. Experiments on Middlebury data sets verify the superior performance of the proposed algorithm, and very effective compromise between the accuracy and computation load is achieved.