1 Hough Circle Transform {#tutorial_hough_circle} 2 ====================== 3 4 Goal 5 ---- 6 7 In this tutorial you will learn how to: 8 9 - Use the OpenCV function @ref cv::HoughCircles to detect circles in an image. 10 11 Theory 12 ------ 13 14 ### Hough Circle Transform 15 16 - The Hough Circle Transform works in a *roughly* analogous way to the Hough Line Transform 17 explained in the previous tutorial. 18 - In the line detection case, a line was defined by two parameters \f$(r, \theta)\f$. In the circle 19 case, we need three parameters to define a circle: 20 21 \f[C : ( x_{center}, y_{center}, r )\f] 22 23 where \f$(x_{center}, y_{center})\f$ define the center position (green point) and \f$r\f$ is the radius, 24 which allows us to completely define a circle, as it can be seen below: 25 26  27 28 - For sake of efficiency, OpenCV implements a detection method slightly trickier than the standard 29 Hough Transform: *The Hough gradient method*, which is made up of two main stages. The first 30 stage involves edge detection and finding the possible circle centers and the second stage finds 31 the best radius for each candidate center. For more details, please check the book *Learning 32 OpenCV* or your favorite Computer Vision bibliography 33 34 Code 35 ---- 36 37 -# **What does this program do?** 38 - Loads an image and blur it to reduce the noise 39 - Applies the *Hough Circle Transform* to the blurred image . 40 - Display the detected circle in a window. 41 42 -# The sample code that we will explain can be downloaded from [here](https://github.com/Itseez/opencv/tree/master/samples/cpp/houghcircles.cpp). 43 A slightly fancier version (which shows trackbars for 44 changing the threshold values) can be found [here](https://github.com/Itseez/opencv/tree/master/samples/cpp/tutorial_code/ImgTrans/HoughCircle_Demo.cpp). 45 @include samples/cpp/houghcircles.cpp 46 47 Explanation 48 ----------- 49 50 -# Load an image 51 @code{.cpp} 52 src = imread( argv[1], 1 ); 53 54 if( !src.data ) 55 { return -1; } 56 @endcode 57 -# Convert it to grayscale: 58 @code{.cpp} 59 cvtColor( src, src_gray, COLOR_BGR2GRAY ); 60 @endcode 61 -# Apply a Gaussian blur to reduce noise and avoid false circle detection: 62 @code{.cpp} 63 GaussianBlur( src_gray, src_gray, Size(9, 9), 2, 2 ); 64 @endcode 65 -# Proceed to apply Hough Circle Transform: 66 @code{.cpp} 67 vector<Vec3f> circles; 68 69 HoughCircles( src_gray, circles, HOUGH_GRADIENT, 1, src_gray.rows/8, 200, 100, 0, 0 ); 70 @endcode 71 with the arguments: 72 73 - *src_gray*: Input image (grayscale). 74 - *circles*: A vector that stores sets of 3 values: \f$x_{c}, y_{c}, r\f$ for each detected 75 circle. 76 - *HOUGH_GRADIENT*: Define the detection method. Currently this is the only one available in 77 OpenCV. 78 - *dp = 1*: The inverse ratio of resolution. 79 - *min_dist = src_gray.rows/8*: Minimum distance between detected centers. 80 - *param_1 = 200*: Upper threshold for the internal Canny edge detector. 81 - *param_2* = 100\*: Threshold for center detection. 82 - *min_radius = 0*: Minimum radio to be detected. If unknown, put zero as default. 83 - *max_radius = 0*: Maximum radius to be detected. If unknown, put zero as default. 84 85 -# Draw the detected circles: 86 @code{.cpp} 87 for( size_t i = 0; i < circles.size(); i++ ) 88 { 89 Point center(cvRound(circles[i][0]), cvRound(circles[i][1])); 90 int radius = cvRound(circles[i][2]); 91 // circle center 92 circle( src, center, 3, Scalar(0,255,0), -1, 8, 0 ); 93 // circle outline 94 circle( src, center, radius, Scalar(0,0,255), 3, 8, 0 ); 95 } 96 @endcode 97 You can see that we will draw the circle(s) on red and the center(s) with a small green dot 98 99 -# Display the detected circle(s): 100 @code{.cpp} 101 namedWindow( "Hough Circle Transform Demo", WINDOW_AUTOSIZE ); 102 imshow( "Hough Circle Transform Demo", src ); 103 @endcode 104 -# Wait for the user to exit the program 105 @code{.cpp} 106 waitKey(0); 107 @endcode 108 109 Result 110 ------ 111 112 The result of running the code above with a test image is shown below: 113 114  115
