Identifying circles on digital images is one of the problems that occur in the perception of objects when processing different images/videos. Solutions are used in many procedures, such as analysis of X-ray images, recognition of circular objects in various videos, robot vision, ...
Identification process takes place from image acquisition to final results that we have obtained with the program. The steps in-between are preparing pictures, drawing objects, various object detection, high-level processing and others. This thesis focuses on algorithms that allow us to fit circles based on specified points by using least-squares method. We present some of the most common approaches, among which we are looking for the most flexible, resilient, quick and simple solutions. All algorithms are analysed, tested and explained with code examples.
Based on the presented tests we can choose an algorithm that is best suited for our program.
|
