| 摘要 |
A system and method for extracting depth information from at least two images employing region-based filtering for reducing artifacts are provided. The present disclosure provides a post-processing algorithm or function for reducing the artifacts generated by scanline Dynamic Programming (DP) or other similar methods. The system and method provides for acquiring a first image and a second image from a scene (202), estimating the disparity of at least one point in the first image with at least one corresponding point in the second image to generate a disparity map (204), segmenting at least one of the first or second images into at least one region (208), and filtering the disparity map based on the segmented regions (210). Furthermore, anisotropic filters are employed, which have a great smoothing effect along the vertical direction than that of the horizontal direction, and therefore, reduce stripe artifacts without significantly blurring the depth boundaries. |
