ptemple's profile - activity

In the code of OpenCV; in the function detect or create for instance

Could you add a check in the code with a more relevant message then please ?

In b = ORB::create(400,0,8,31,0,2,0,31,20);

the second argument should be a float greater or equals to 1 according to the documentation at http://docs.opencv.org/trunk/db/d95/classcv_1_1ORB.html (http://docs.opencv.org/trunk/db/d95/c...)

while I set it to 0

When using your code your image given twice (for both arguments); the execution reveals some assertion failure for different descriptors : for AKAZE-DESCRIPTOR_KAZE_UPRIGHT and the hamming distance; your usage of ORB seems ok though; my configuration given by ORB::create() is as follows :

b = ORB::create(400,0,8,31,0,2,0,31,20);

the last 0 should be an index to something but I don't know which are available so... and my configuration leverage some assertion failure

OpenCV error : Assertion failed (s >= 0) in setSize, matrix.cpp, line 297

Why do I get the assertion then ? Are there any requirements on the value given to the ORB::create() function (like min/max boundary values) that would trigger the assertion ?

Yes, sorry, I forgot this line; I edited the post, I use some values from a configuration file to create the structure.

Hi there, I'm trying to use ORB points of interest to try to retrieve an object into an image (compute points of interest; then associated descriptors and finally try to see if we can pair them and retrieve a transformation).

Here is a part of the code I use :

Ptr<cv::ORB> descriptor_ORB;
descriptor_ORB = cv::ORB::create(/*some values from a configuration file with atoi() and atof() to format values*/);

Mat model = imread(filename); //string filename -> the name of an image file
std::vector<KeyPoint> keypoints_model;
cout << "size model : " << model.cols << "\t" << models.rows << endl; //display 123 and 62 -> successfully load
Mat gray_model = model;
cvtColor(gray_model, gray_model, COLOR_BGR2GRAY); // convert to grayscale as in examples
descriptor_ORB -> detect (gray_model, keypoints_model); // compute ORB points of interest on grayscale image and put the points in the keypoints_model vector

This last line gives me an assertion failure :

OpenCV Error : Assertion failed (0 <= roi.x && 0 <= roi.width && roi.x +roi.width <= m.cols && 0 <= roi.y && 0 <= roi.height && roi.y + roi.height <= m.rows) in Mat, file Matrix.cpp, line 499

Is it normal ? It is likely that the detect function is unable to retrieve any points of interest...

To avoid early, termination I put the detect call in a try/catch statement.