This is a black and white image that I created in Paint Shop by extracting all the edges out of a pastoral image my dad took in Delaware over Christmas.
This is the result of applying the Hough Line Transform on the above image. Pretty, isn’t it? What you are seeing is all the possible polar parameterizations (within the windows of approximation) of all the lines in the previous image. The brighter a particular pixel is, the more strongly that line is shown in the original. This particular output image was post-processed in Paint Shop to increase the gamma in the intensity histogram, so the sense of there being peaks has been destroyed, in the name of creating a more aesthetically pleasing image. Normally, peaks in the output image would correspond to the strongest lines found in the source image. So, for example, if source image contained only the line y = 5, there would be a single peak in the output image, at (5, PI/2).
I wrote a Hough Transformer in about an hour yesterday to solve a homework problem for CS223B. I’ve uploaded the source to my code library, if you are interested (though I will admit it is rather arcane).
