Characterisation of circular objects based on the hough transform

This paper presents a new image analysis approach based on the Hough Transform, for the characterisation of natural objects of approximately circular shape in real situations. To accomplish this purpose, an innovative interpretation and filtering methodology of the three dimensional Hough Transform (HT) accumulation space for circles was conceived and implemented, which proved to be crucial when dealing with shape and texture irregularities of natural objects. Some results that demonstrate the robustness of the developed approach under real testing conditions are also presented, namely concerning its application into a vision system that automates the volume measurement of piles of timber logs, allowing the evaluation of their diameter distributions.