Convex Polygon Detection using Linear Hough Transform

Polygon detection framework for bitmap images written in C that implements Otsu's thresholding and a Deriche edge detector for blurring, calculation of magnitude and gradient direction, non-maximum suppression and hysteresis thresholding. As a final step, it performs a Hough transform to detect the number and orientation of each polygon edge in the source image.

For a detailed overview, be sure to checkout the guide at the following link: Deriche-Edges-HoughTransform.

Building

Clone the repo and run the following commands in shell, which should output an executable named edgedetect :

./configure && make

Running Instructions

Once built, you can execute edgedetect like so:

./edgedetect examples/image1.bmp

Optional arguments

--verbose            :
--write-csv          : Writes a CSV file for every intermediate step
--set-blur-alpha     : Blur alpha, greater than 0.0
--set-gradient-alpha : Gradient alpha, greater than 0.0

Example usage

./edgedetect examples/diamond.bmp --set-blur-alpha 1.0 --set-gradient-alpha 10.0