Skip to content

Latest commit

 

History

History
51 lines (26 loc) · 1.44 KB

README.md

File metadata and controls

51 lines (26 loc) · 1.44 KB
Raw

stereo-disparity-map

Computes a disparity map from a stereo pair of black and white images.

Use

PIXEL_DISP(leftImage, rightImage, outputImage, plot, greyscale, scan_window_size, support_scale);

e.g.

PIXEL_DISP('IN/pentagon_right.bmp', 'IN/pentagon_left.bmp', 'OUT/Pentagon_7_3', 0, 1, 7, 3);

Example Stereo Pair & Resulting Disparity Map

Disparity Map Disparity Map Disparity Map

Algorithm Summary

read in L and R images and if necessary, convert both to greyscale

  loop for each pixel in the left image

      generate matrix of left scan window for the current pixel

      calculate boundaries for the large right support window

      generate matrix of large right support window

      begin loop for each pixel in the large right support window

          get matrix of scan window size for the current pixel in support window

          calculate disparity by performing sum of squared differences

          if the disparity for the current right scan window is closer to 0 then

          note disparity and x value within the right support window

      end loop

      calculate the vector from the current row point to closest disparity match

      map the value between 0-255 for display purposes

      add value to array of disparity map array

  end loop

  write image out

end