An Algorithm to Recognize and Classify Circular Objects from Image on Basis of Their Radius

Abstract

For the computer vision, fast and accurate detection of an object is challenging. Detecting a circular object in a cluttered image has always been a problem. Circular object detections has wide applications in the field of biometrics, automobile and other mechanical production industries. The traditional existing circular object detection are maximum likelihood estimation (MLE) and voting-based methods. The voting based methods have high memory requirements and more computational complexity while these are less sensitive to noise. MLE approach consumes less memory and are efficient in terms of computational complexity but these approaches are more prone to noise. This paper proposes modified Hough transform based algorithm for detection of circular objects within other shaped objects also it can identify circular objects on basis of diameter. The proposed algorithm worked efficiently and detected the circular objects on basis of diameters with very less computational time and less memory consumption.