Stereo Matching using SGM based on python

Introduction

This project is trying to implementing stereo matching by python based on the middleburry dataset. The whole process contains the following steps.

• Census Transform
• cost volume Calculation by hamming distance
• cost aggregation
• WTA to calculate the disparity
• consistency check to exclude bad points
• map the disparity to 0~255 and display

Environment

• Python 3.7
• NumPy
• OpenCV

File structure

• dataset: some standard images contained in middleburry dataset
• SGM_fun.py: all the source codes