IDENTIFYING GEOMETRICAL OBJECTS USING IMAGE ANALYSIS

One of the major challenges in computer vision is to extract the objects automatically from an image. This work seeks to detect objects in image. Herein, the objects are supposed to be circles. This paper presents an algorithm to determine the main characteristics to define the circles in image. The main ingredients of the algorithm are thresholding, scale-space and median filter methods. The identification is made by estimate the three characteristics which are number, sizes and locations of the object. A proposed algorithm has been applied to synthetic and real datasets and yields good results.