This is the code in paper Efficient Methods Using Slanted Windows for Slanted Surfaces, which I pulished when I did research on stereo matching during master study.
The cost computation: absolute difference (AD) + census transform (CT).
The cost aggregation: bilateral filter is used to adapte weight according the slanted support window in disparity space.
The post processing: outlier detection, outlier filling and refinement of disparities.
The code is a Visual Studio 2010 project on Windows x64 platform. To build the project, you need to configure OpenCV. (>=version 2.4.6). The code requires no platform-dependent libraries. Thus, it is easy to compile it on other platforms with OpenCV.
Citation is very important for researchers. If you find this code useful, please cite:
@inproceedings{stereo_matching_slanted_support,
author = {Xuesong LI and Jianguo Liu and Guang Chen and Heng Fu},
title = {Efficient Methods Using Slanted Support Windows for Slanted Surfaces},
journal = {IET Computer Vision},
year = 2016,
pages = {384-391},
month = 8,
note = {http://apps.webofknowledge.com/Search.do?product=WOS&SID=C5oqWs1CGyfqY8n42RB&search_mode=GeneralSearch&prID=e0079d24-9970-4978-b331-45a7e5e80791},
volume = 10
}