The present report outlines the design, implementation and performance of the application of Hough

Transform (HT) to detect circles in arbitrary pictures.

Detecting circles in arbitrary pictures involves a two step process in this project. First, edges within the

image have to be detected then, by means of a voting mechanism the most probable circles will be

located.

Edge detection goal is to look for boundary locations that naturally happen between objects. Objects

normally have continuous intensity values therefore sudden changes in that pattern might indicate a

boundary condition. Changes of intensity in one direction can be calculated by the gradient operator

however, noise can alter meaningful edges. Convolution is then normally applied to account for a small

amount of smoothing, thus reducing noise.-The present report outlines the design, implementation and performance of the application of Hough

Transform (HT) to detect circles in arbitrary pictures.

Detecting circles in arbitrary pictures involves a two step process in this project. First, edges within the

image have to be detected　then, by means of a voting mechanism the most probable circles will be

located.

Edge detection goal is to look for boundary locations that naturally happen between objects. Objects

normally have continuous intensity values therefore sudden changes in that pattern might indicate a

boundary condition. Changes of intensity in one direction can be calculated by the gradient operator

however, noise can alter meaningful edges. Convolution is then normally applied to account for a small

amount of smoothing, thus reducing noise.