Stable high-performance circle feature detection method

Abstract

The invention discloses a stable high-performance circle feature detection method. As for different image qualities and speeds, a gradient direction-based method or a radial symmetry-based method needs to be adopted respectively to determine coordinates of a circle center. In the gradient direction-based method, through solving the maximum value of an energy function, the coordinates of the circle center are determined, a gradient ascent method is used for quickening the calculation speed, and a surface fitting method is used to solve subpixel-precision circle center coordinates. As for a low-quality picture, a circle center-based gray symmetry features are used, the energy function is determined through calculating the pixel symmetry with the circle center as a starting point, and the circle center coordinates exist at the position with the minimum value of the energy function. After the circle center subpixel coordinates are determined, through solving points with the maximum gradient values in a certain range in the circle radius direction, the distances between the points and the circle are calculated, and the radius solving process is converted to a process of solving the minimum value of a quadratic equation. As for circle feature extraction, the method of the invention has extremely good stability, accuracy and practicability.

Description

Technical field [0001] The invention relates to the technical field of image recognition, in particular to a method for detecting circle features. Background technique [0002] The most classic methods of image-based circle feature detection are methods based on Hough transform, generalized Hough transform, and random Hough transform. Although the above methods can accurately detect circle features, they have a large amount of calculation and storage space. The low computational efficiency makes this type of method not applicable to real-time systems such as machine vision. After literature search, a large number of circle detection methods are implemented based on Hough transform and evolutionary algorithms, such as the "Fast Circle Detection Algorithm Based on Ordered Hough Transform" proposed by Ye Feng et al. (see Non-Patent Document 1: Ye Feng et al., Fast circle detection algorithm based on ordered Hough transform, Optics and Precision Engineering, 2014 Issue 4), Chen Ke e...

AI Technical Summary

Problems solved by technology

However, the one-dimensional or two-dimensional circle edge point detection method has outliers in the detected edge points due to noise or poor image quality. Usually, RANSAC and other methods are used to remove outliers and retain circle edge points for circle equation fitting.

The problem with this method is that the algorithm is less stable

Patent Document 12 uses a circle to detect the center of the circle based on the symmetry properties of the pixels in the radial direction. Due to the poor image quality in industrial applications, the circular structure is partially missing, which leads to the use of the symmetry properties of the pixels in the radial direction of the circle to detect the center of the circle. The coordinates are more difficult, and the information used in calculating the symmetry is the pixel accumulation value from the detection point to the center of the circle, so it is more difficult to detect the coordinates of the center of the ring structure

[0005] The detection method using the voting mechanism is difficult to strike a balance between the detection speed and the stability of the method

Although the circle detection method using global optimization has good accuracy, its computational efficiency is not suitable for real-time application of machine vision

Images