Research On The Improvement And Application Of Canny Operator In Circle Detection

Edge detection is the basis of image feature extraction,and it is also an indispensable part in image analysis.At present,the operators used in edge detection are mainly divided into first order differential operators,two order differential operators and optimal operators.Among them,the Canny operator in the best operator has the best effect in edge detection,and it is also the most widely applied.Although the Canny operator has a good effect in edge detection,Canny operator still has two flaws.First,The traditional Canny operator uses Gauss filter to smooth the noise,while Gauss filter can remove the noise,but it will lose some edge information while removing the noise.Second,Canny operator need to manually determine thresholds,which may lead to inaccurate threshold estimation,while incorrect threshold estimation leads to poor edge connectivity or false edges.Therefore,it is particularly important to improve the method of Canny operator filtering and the selection of threshold.A binary image of edge detection is obtained by processing Canny operator,and image feature extraction usually needs to be processed on the basis of edge detection.The most common one is circular detection.Circle detection is used in all aspects of life and engineering.One of the most common applications is the pattern recognition and localization of circular devices or icons,and Hough transform is one of the main methods of traditional detection of circles.The Hough transform uses the mapping of the image space to the parameter space to detect the circle.Although the algorithm can detect the circle,its one to many mapping relationships lead to too much computation and take up a large number memory space,at the same time,all kinds of parameters,such as round radius,need to be determined artificially,which leads to the low accuracy of detection.Concerning the above-mentioned issues,views and analyses,which have to do with the preprocessing of edge detection and the follow-up process of edge detection,are peculiarly raised in this paper.and the filtering method and the threshold selection method in the process of preprocessing are improved.For the edge detection follow-up,this paper mainly focuses on the binary detection of the binary image obtained by the edge detection,and proposes a circular detection algorithm based on the connected area labeling algorithm,the main results of this paper are the following two points:(1)In order to overcome the shortcomings of the traditional Canny algorithm,which is easy to lose the edge details of the edge detection,and is prone to a large number of false edges and the threshold needs to be manually determined,an adaptive Canny algorithm based on Dual-domain filtering is proposed.Firstly,the Dual-domain filter is used to instead of the traditional algorithm,which is used to remove the image noise while preserving the image edge information,and more effective than the traditional Canny algorithm;Secondly,the Otsu algorithm is used to calculate the high and low threshold of Canny operator in order to achieve the purpose of high automation.The experimental results show that the improved algorithm can detect more effective edges of the image,but also has a strong adaptability.(2)In order to overcome the shortcomings of traditional Hough transform in the circle detection,such as excessive computation,low precision of detecting concentric circles and low degree of automation,a circle detection algorithm based on connected region labeling algorithm is proposed.This new algorithm first solves the problem of excessive calculation which the traditional algorithm to process the image for the circle.Second,confirming the center and radius of the circle according to the characteristics,this avoids the detection problem of concentric circles.Finally,for the aim of automation,the 8 neighborhood pixels of the center of the circle were taken to determine the best and final circle.The experimental results show that the algorithm mentioned above can detect the circle accurately.Meanwhile,it's less computational and more automated than the Hough transform.