Detection of ChArUco Boards

Prev Tutorial: Detection of ArUco boards
Next Tutorial: Detection of Diamond Markers
ArUco markers and boards are very useful due to their fast detection and their versatility. However, one of the problems of ArUco markers is that the accuracy of their corner positions is not too high, even after applying subpixel refinement.

On the contrary, the corners of chessboard patterns can be refined more accurately since each corner is surrounded by two black squares. However, finding a chessboard pattern is not as versatile as finding an ArUco board: it has to be completely visible and occlusions are not permitted.

A ChArUco board tries to combine the benefits of these two approaches:

Charuco definition

The ArUco part is used to interpolate the position of the chessboard corners, so that it has the versatility of marker boards, since it allows occlusions or partial views. Moreover, since the interpolated corners belong to a chessboard, they are very accurate in terms of subpixel accuracy.

When high precision is necessary, such as in camera calibration, Charuco boards are a better option than standard ArUco boards.

Goal

In this tutorial you will learn:

• How to create a charuco board ?
• How to detect the charuco corners without performing camera calibration ?
• How to detect the charuco corners with camera calibration and pose estimation ?

Source code

You can find this code in samples/cpp/tutorial_code/objectDetection/detect_board_charuco.cpp

Here's a sample code of how to achieve all the stuff enumerated at the goal list.

 int squaresX = parser.get<int>("w");
 int squaresY = parser.get<int>("h");
 float squareLength = parser.get<float>("sl");
 float markerLength = parser.get<float>("ml");
 bool refine = parser.has("rs");
 int camId = parser.get<int>("ci");
 
 string video;
 if(parser.has("v")) {
 video = parser.get<string>("v");
 }
 
 Mat camMatrix, distCoeffs;
 readCameraParamsFromCommandLine(parser, camMatrix, distCoeffs);
 aruco::DetectorParameters detectorParams = readDetectorParamsFromCommandLine(parser);
 aruco::Dictionary dictionary = readDictionatyFromCommandLine(parser);
 
 if(!parser.check()) {
 parser.printErrors();
 return 0;
 }
 
 VideoCapture inputVideo;
 int waitTime = 0;
 if(!video.empty()) {
 inputVideo.open(video);
 } else {
 inputVideo.open(camId);
 waitTime = 10;
 }
 
 float axisLength = 0.5f * ((float)min(squaresX, squaresY) * (squareLength));
 
 // create charuco board object
 aruco::CharucoBoard charucoBoard(Size(squaresX, squaresY), squareLength, markerLength, dictionary);
 
 // create charuco detector
 aruco::CharucoParameters charucoParams;
 charucoParams.tryRefineMarkers = refine; // if tryRefineMarkers, refineDetectedMarkers() will be used in detectBoard()
 charucoParams.cameraMatrix = camMatrix; // cameraMatrix can be used in detectBoard()
 charucoParams.distCoeffs = distCoeffs; // distCoeffs can be used in detectBoard()
 aruco::CharucoDetector charucoDetector(charucoBoard, charucoParams, detectorParams);
 
 double totalTime = 0;
 int totalIterations = 0;
 
 while(inputVideo.grab()) {
 Mat image, imageCopy;
 inputVideo.retrieve(image);
 
 double tick = (double)getTickCount();
 
 vector<int> markerIds, charucoIds;
 vector<vector<Point2f> > markerCorners;
 vector<Point2f> charucoCorners;
 Vec3d rvec, tvec;
 
 // detect markers and charuco corners
 charucoDetector.detectBoard(image, charucoCorners, charucoIds, markerCorners, markerIds);
 
 // estimate charuco board pose
 bool validPose = false;
 if(camMatrix.total() != 0 && distCoeffs.total() != 0 && charucoIds.size() >= 4) {
 Mat objPoints, imgPoints;
 charucoBoard.matchImagePoints(charucoCorners, charucoIds, objPoints, imgPoints);
 validPose = solvePnP(objPoints, imgPoints, camMatrix, distCoeffs, rvec, tvec);
 }
 
 double currentTime = ((double)getTickCount() - tick) / getTickFrequency();
 totalTime += currentTime;
 totalIterations++;
 if(totalIterations % 30 == 0) {
 cout << "Detection Time = " << currentTime * 1000 << " ms "
 << "(Mean = " << 1000 * totalTime / double(totalIterations) << " ms)" << endl;
 }
 
 // draw results
 image.copyTo(imageCopy);
 if(markerIds.size() > 0) {
 aruco::drawDetectedMarkers(imageCopy, markerCorners);
 }
 
 if(charucoIds.size() > 0) {
 aruco::drawDetectedCornersCharuco(imageCopy, charucoCorners, charucoIds, cv::Scalar(255, 0, 0));
 }
 
 if(validPose)
 cv::drawFrameAxes(imageCopy, camMatrix, distCoeffs, rvec, tvec, axisLength);
 
 imshow("out", imageCopy);
 if(waitKey(waitTime) == 27) break;
 }

ChArUco Board Creation

The aruco module provides the cv::aruco::CharucoBoard class that represents a Charuco Board and which inherits from the cv::aruco::Board class.

This class, as the rest of ChArUco functionalities, are defined in:

#include <opencv2/objdetect/charuco_detector.hpp>

To define a cv::aruco::CharucoBoard, it is necessary:

• Number of chessboard squares in X and Y directions.
• Length of square side.
• Length of marker side.
• The dictionary of the markers.
• Ids of all the markers.

As for the cv::aruco::GridBoard objects, the aruco module provides to create cv::aruco::CharucoBoard easily. This object can be easily created from these parameters using the cv::aruco::CharucoBoard constructor:

 aruco::Dictionary dictionary = readDictionatyFromCommandLine(parser);
 cv::aruco::CharucoBoard board(Size(squaresX, squaresY), (float)squareLength, (float)markerLength, dictionary);

• The first parameter is the number of squares in X and Y direction respectively.
• The second and third parameters are the length of the squares and the markers respectively. They can be provided in any unit, having in mind that the estimated pose for this board would be measured in the same units (usually meters are used).
• Finally, the dictionary of the markers is provided.

The ids of each of the markers are assigned by default in ascending order and starting on 0, like in cv::aruco::GridBoard constructor. This can be easily customized by accessing to the ids vector through board.ids, like in the cv::aruco::Board parent class.

Once we have our cv::aruco::CharucoBoard object, we can create an image to print it. There are two ways to do this:

1. By using the script doc/patter_tools/gen_pattern.py, see Create calibration pattern.
2. By using the function cv::aruco::CharucoBoard::generateImage().

The function cv::aruco::CharucoBoard::generateImage() is provided in cv::aruco::CharucoBoard class and can be called by using the following code:

 Mat boardImage;
 Size imageSize;
 imageSize.width = squaresX * squareLength + 2 * margins;
 imageSize.height = squaresY * squareLength + 2 * margins;
 board.generateImage(imageSize, boardImage, margins, borderBits);

• The first parameter is the size of the output image in pixels. If this is not proportional to the board dimensions, it will be centered on the image.
• The second parameter is the output image with the charuco board.
• The third parameter is the (optional) margin in pixels, so none of the markers are touching the image border.
• Finally, the size of the marker border, similarly to cv::aruco::generateImageMarker() function. The default value is 1.

The output image will be something like this:

A full working example is included in the create_board_charuco.cpp inside the samples/cpp/tutorial_code/objectDetection/.

The samples create_board_charuco.cpp now take input via commandline via the cv::CommandLineParser. For this file the example parameters will look like:

"_output_path_/chboard.png" -w=5 -h=7 -sl=100 -ml=60 -d=10

ChArUco Board Detection

When you detect a ChArUco board, what you are actually detecting is each of the chessboard corners of the board.

Each corner on a ChArUco board has a unique identifier (id) assigned. These ids go from 0 to the total number of corners in the board. The steps of charuco board detection can be broken down to the following steps:

• Taking input Image

 Mat image, imageCopy;
 inputVideo.retrieve(image);

The original image where the markers are to be detected. The image is necessary to perform subpixel refinement in the ChArUco corners.

• Reading the camera calibration Parameters(only for detection with camera calibration)

 if(parser.has("c")) {
 bool readOk = readCameraParameters(parser.get<std::string>("c"), camMatrix, distCoeffs);
 if(!readOk) {
 throw std::runtime_error("Invalid camera file\n");
 }
 }

The parameters of readCameraParameters are:

• The first parameter is the path to the camera intrinsic matrix and distortion coefficients.
• The second and third parameters are cameraMatrix and distCoeffs.

This function takes these parameters as input and returns a boolean value of whether the camera calibration parameters are valid or not. For detection of charuco corners without calibration, this step is not required.

• Detecting the markers and interpolation of charuco corners from markers

The detection of the ChArUco corners is based on the previous detected markers. So that, first markers are detected, and then ChArUco corners are interpolated from markers. The method that detect the ChArUco corners is cv::aruco::CharucoDetector::detectBoard().

 // detect markers and charuco corners
 charucoDetector.detectBoard(image, charucoCorners, charucoIds, markerCorners, markerIds);

The parameters of detectBoard are:

• image - Input image.
• charucoCorners - output list of image positions of the detected corners.
• charucoIds - output ids for each of the detected corners in charucoCorners.
• markerCorners - input/output vector of detected marker corners.
• markerIds - input/output vector of identifiers of the detected markers

If markerCorners and markerIds are empty, the function will detect aruco markers and ids.

If calibration parameters are provided, the ChArUco corners are interpolated by, first, estimating a rough pose from the ArUco markers and, then, reprojecting the ChArUco corners back to the image.

On the other hand, if calibration parameters are not provided, the ChArUco corners are interpolated by calculating the corresponding homography between the ChArUco plane and the ChArUco image projection.

The main problem of using homography is that the interpolation is more sensible to image distortion. Actually, the homography is only performed using the closest markers of each ChArUco corner to reduce the effect of distortion.

When detecting markers for ChArUco boards, and specially when using homography, it is recommended to disable the corner refinement of markers. The reason of this is that, due to the proximity of the chessboard squares, the subpixel process can produce important deviations in the corner positions and these deviations are propagated to the ChArUco corner interpolation, producing poor results.

Note

To avoid deviations, the margin between chessboard square and aruco marker should be greater than 70% of one marker module.

Furthermore, only those corners whose two surrounding markers have be found are returned. If any of the two surrounding markers has not been detected, this usually means that there is some occlusion or the image quality is not good in that zone. In any case, it is preferable not to consider that corner, since what we want is to be sure that the interpolated ChArUco corners are very accurate.

After the ChArUco corners have been interpolated, a subpixel refinement is performed.

Once we have interpolated the ChArUco corners, we would probably want to draw them to see if their detections are correct. This can be easily done using the cv::aruco::drawDetectedCornersCharuco() function:

 aruco::drawDetectedCornersCharuco(imageCopy, charucoCorners, charucoIds, cv::Scalar(255, 0, 0));

• imageCopy is the image where the corners will be drawn (it will normally be the same image where the corners were detected).
• The outputImage will be a clone of inputImage with the corners drawn.
• charucoCorners and charucoIds are the detected Charuco corners from the cv::aruco::CharucoDetector::detectBoard() function.
• Finally, the last parameter is the (optional) color we want to draw the corners with, of type cv::Scalar.

For this image:

Image with Charuco board

The result will be:

Charuco board detected

In the presence of occlusion. like in the following image, although some corners are clearly visible, not all their surrounding markers have been detected due occlusion and, thus, they are not interpolated:

Charuco detection with occlusion

Sample video:

A full working example is included in the detect_board_charuco.cpp inside the samples/cpp/tutorial_code/objectDetection/.

The samples detect_board_charuco.cpp now take input via commandline via the cv::CommandLineParser. For this file the example parameters will look like:

-w=5 -h=7 -sl=0.04 -ml=0.02 -d=10 -v=/path_to_opencv/opencv/doc/tutorials/objdetect/charuco_detection/images/choriginal.jpg

ChArUco Pose Estimation

The final goal of the ChArUco boards is finding corners very accurately for a high precision calibration or pose estimation.

The aruco module provides a function to perform ChArUco pose estimation easily. As in the cv::aruco::GridBoard, the coordinate system of the cv::aruco::CharucoBoard is placed in the board plane with the Z axis pointing in, and centered in the bottom left corner of the board.

Note

After OpenCV 4.6.0, there was an incompatible change in the coordinate systems of the boards, now the coordinate systems are placed in the boards plane with the Z axis pointing in the plane (previously the axis pointed out the plane). objPoints in CW order correspond to the Z-axis pointing in the plane. objPoints in CCW order correspond to the Z-axis pointing out the plane. See PR https://github.com/opencv/opencv_contrib/pull/3174

To perform pose estimation for charuco boards, you should use cv::aruco::CharucoBoard::matchImagePoints() and cv::solvePnP():

 // estimate charuco board pose
 bool validPose = false;
 if(camMatrix.total() != 0 && distCoeffs.total() != 0 && charucoIds.size() >= 4) {
 Mat objPoints, imgPoints;
 charucoBoard.matchImagePoints(charucoCorners, charucoIds, objPoints, imgPoints);
 validPose = solvePnP(objPoints, imgPoints, camMatrix, distCoeffs, rvec, tvec);
 }

• The charucoCorners and charucoIds parameters are the detected charuco corners from the cv::aruco::CharucoDetector::detectBoard() function.
• The cameraMatrix and distCoeffs are the camera calibration parameters which are necessary for pose estimation.
• Finally, the rvec and tvec parameters are the output pose of the Charuco Board.
• cv::solvePnP() returns true if the pose was correctly estimated and false otherwise. The main reason of failing is that there are not enough corners for pose estimation or they are in the same line.

The axis can be drawn using cv::drawFrameAxes() to check the pose is correctly estimated. The result would be: (X:red, Y:green, Z:blue)

Charuco Board Axis

A full working example is included in the detect_board_charuco.cpp inside the samples/cpp/tutorial_code/objectDetection/.

The samples detect_board_charuco.cpp now take input via commandline via the cv::CommandLineParser. For this file the example parameters will look like:

-w=5 -h=7 -sl=0.04 -ml=0.02 -d=10
-v=/path_to_opencv/opencv/doc/tutorials/objdetect/charuco_detection/images/choriginal.jpg
-c=/path_to_opencv/opencv/samples/cpp/tutorial_code/objectDetection/tutorial_camera_charuco.yml