Fast Positioning Method for Concentric Circle Marks Based on Color and Shape Features

Abstract

The most classic method for finding a circle mark in an image and positioning the center of the circle is the Hough circle detection. However, this method has many problems, such as large amount of computation, low detection accuracy and large memory usage, which cannot meet the requirement of real-time performance. This paper proposes a positioning method based on color and shape features, first by color segmentation area of interest, then use median filter to filter the noise in image, using edge detection and contour to find the outline of the image, finally from the view of the mathematical description of circular shape characteristics, the contour filter, so as to quickly find the center of concentric circles. Experiments show that compared with the improved Hough gradient method, the proposed method has lower complexity, faster calculation speed and can find the marker area more effectively, especially suitable for real-time requirements such as drone vision guidance and accurate landing.