/StereoVision

First-principles implementation of stereo depth construction. Produces a real-world 3D point cloud that corresponds to the scene depicted in the 2 given stereo images.

Primary LanguagePython

StereoVision

First-principles implementation of stereo depth construction. 3D-vision is achieved using 2D camera images.

Project includes both a Python and an MATLAB version, both produces a real-world 3D point cloud that corresponds to the scene depicted in the 2 given stereo images.

  1. Block Matching method using Sum of Absolute Difference (SAD) is implemented in both Python and MATLAB.
  2. Integral Images method is implemented in Python.

The dataset tested on is found at: https://vision.middlebury.edu/stereo/data/scenes2014/