CODEWITHSUNDEEP
c++opencv
L7ColWinters
L7ColWinters

Reputation: 1362

OpenCV Mat format

read the image in

Mat img=imread("i000qa-fn.jpg",CV_LOAD_IMAGE_COLOR);

try to find objects...

faces = cvHaarDetectObjects(img,cascade, storage, 1.2, 2, CV_HAAR_DO_CANNY_PRUNING, Size(0, 0));

and walla...

OpenCV Error: Bad flag (parameter or structure field) (Unrecognized or unsupported array type) in cvGetMat, file /media/Data/sdks/OpenCV-2.4.2/modules/core/src/array.cpp, line 2482

when I do imshow, the image is there as it should be.

Upvotes: 2

Views: 4565

Answers (2)

Andrey Kamaev
Andrey Kamaev

Reputation: 30152

cvHaarDetectObjects expects IplImage or CvMat but you are passing cv::Mat object.

So you need a conversion like this:

IplImage img1 = img;
faces = cvHaarDetectObjects(&img1, cascade, storage, 1.2, 2, CV_HAAR_DO_CANNY_PRUNING, Size(0, 0));

Upvotes: 3

Barney Szabolcs
Barney Szabolcs

Reputation: 12554

No, Andrey (@AndreyKamaev), you need a different function instead:

#include <opencv2/core/core.hpp>
#include <opencv2/objdetect/objdetect.hpp>
using namespace cv;

Mat img = imread(img_path);
CascadeClassifier haar_cascade.load(path);

vector<Rect> detection_rois;
haar_cascade.detectMultiScale(img, detection_rois, 1.2, 2, 
                                  0|CV_HAAR_DO_CANNY_PRUNING);

That is how Haar detector is used in C++ since Opencv 2.3.1 that is since Aug. 2011. Also let me attach a documentation.

Here is a proof, below. :) I did a Haar_detector wrapper around this cv::CascadeClassifier -- which is really an Adaboost cascade classifier with Haar-like features, therefore the name.

The proof

Upvotes: 2

Related Questions