I have no problem drawing a texture to the screen but I cant get the right pixels when printing them from memory. I have a 4x4 png image with 4 black pixels and I am trying to print them. This is what I do:
glBindTexture(GL_TEXTURE_2D, m_textureId);
const int size = m_width * m_height * 4;
GLubyte pixels[size];
glReadPixels(0, 0, m_width, m_height, GL_RGBA, GL_UNSIGNED_BYTE, pixels);
if(glGetError() != GL_NO_ERROR)
assert(false && "opengl error");
for(int index = 0; index < size; index+=4)
{
cout << "red " << (unsigned)pixels[index+0] << endl;
cout << "green " << (unsigned)pixels[index+1] << endl;
cout << "blue " << (unsigned)pixels[index+2] << endl;
}
But I get all random values and not the one I expect. Can anyone see what I am doing wrong?
glReadPixels reads from framebuffers, not textures. To retrieve the contents of a texture object use glGetTexImage: http://www.opengl.org/sdk/docs/man/xhtml/glGetTexImage.xml
Related
As mentioned in the title, I have part of my vscode Chinese, which is troublesome since looking up error codes in English will be easier. How can I turn it into complete English? I've tried How to set Visual Studio Code Terminal output to English and it didn't work for me.
Since someone asked for the code to be post in text, this is the code (I'm using a tutorial code for camera calibration from opencv, so the error code appearing also confused me):
#include <iostream>
#include <opencv2/calib3d.hpp>
#include <opencv2/core.hpp>
#include <opencv2/highgui.hpp>
#include <opencv2/imgproc.hpp>
int main(int argc, char **argv) {
(void)argc;
(void)argv;
std::vector<cv::String> fileNames;
cv::glob("../calibration/Image*.png", fileNames, false);
cv::Size patternSize(25 - 1, 18 - 1);
std::vector<std::vector<cv::Point2f>> q(fileNames.size());
std::vector<std::vector<cv::Point3f>> Q;
// 1. Generate checkerboard (world) coordinates Q. The board has 25 x 18
// fields with a size of 15x15mm
int checkerBoard[2] = {25,18};
// Defining the world coordinates for 3D points
std::vector<cv::Point3f> objp;
for(int i = 1; i<checkerBoard[1]; i++){
for(int j = 1; j<checkerBoard[0]; j++){
objp.push_back(cv::Point3f(j,i,0));
}
}
std::vector<cv::Point2f> imgPoint;
// Detect feature points
std::size_t i = 0;
for (auto const &f : fileNames) {
std::cout << std::string(f) << std::endl;
// 2. Read in the image an call cv::findChessboardCorners()
cv::Mat img = cv::imread(fileNames[i]);
cv::Mat gray;
cv::cvtColor(img, gray, cv::COLOR_RGB2GRAY);
bool patternFound = cv::findChessboardCorners(gray, patternSize, q[i], cv::CALIB_CB_ADAPTIVE_THRESH + cv::CALIB_CB_NORMALIZE_IMAGE + cv::CALIB_CB_FAST_CHECK);
// 2. Use cv::cornerSubPix() to refine the found corner detections
if(patternFound){
cv::cornerSubPix(gray, q[i],cv::Size(11,11), cv::Size(-1,-1), cv::TermCriteria(cv::TermCriteria::EPS + cv::TermCriteria::MAX_ITER, 30, 0.1));
Q.push_back(objp);
}
// Display
cv::drawChessboardCorners(img, patternSize, q[i], patternFound);
cv::imshow("chessboard detection", img);
cv::waitKey(0);
i++;
}
cv::Matx33f K(cv::Matx33f::eye()); // intrinsic camera matrix
cv::Vec<float, 5> k(0, 0, 0, 0, 0); // distortion coefficients
std::vector<cv::Mat> rvecs, tvecs;
std::vector<double> stdIntrinsics, stdExtrinsics, perViewErrors;
int flags = cv::CALIB_FIX_ASPECT_RATIO + cv::CALIB_FIX_K3 +
cv::CALIB_ZERO_TANGENT_DIST + cv::CALIB_FIX_PRINCIPAL_POINT;
cv::Size frameSize(1440, 1080);
std::cout << "Calibrating..." << std::endl;
// 4. Call "float error = cv::calibrateCamera()" with the input coordinates
// and output parameters as declared above...
float error = cv::calibrateCamera(Q, q, frameSize, K, k, rvecs, tvecs, flags);
std::cout << "Reprojection error = " << error << "\nK =\n"
<< K << "\nk=\n"
<< k << std::endl;
// Precompute lens correction interpolation
cv::Mat mapX, mapY;
cv::initUndistortRectifyMap(K, k, cv::Matx33f::eye(), K, frameSize, CV_32FC1,
mapX, mapY);
// Show lens corrected images
for (auto const &f : fileNames) {
std::cout << std::string(f) << std::endl;
cv::Mat img = cv::imread(f, cv::IMREAD_COLOR);
cv::Mat imgUndistorted;
// 5. Remap the image using the precomputed interpolation maps.
cv::remap(img, imgUndistorted, mapX, mapY, cv::INTER_LINEAR);
// Display
cv::imshow("undistorted image", imgUndistorted);
cv::waitKey(0);
}
return 0;
}
boost::icl::interval_map<int, boost::icl::interval_map<int, boost::icl::interval_set<int>>> larger, smaller;
larger.add(make_pair(boost::icl::discrete_interval<int>::closed(102,104),
boost::icl::interval_map<int, boost::icl::interval_set<int>>{
make_pair(boost::icl::discrete_interval<int>::closed(0,0), boost::icl::interval_set<int {boost::icl::discrete_interval<int>::closed(2,4)})}));
smaller.add(make_pair(boost::icl::discrete_interval<int>::closed(103,103),
boost::icl::interval_map<int, boost::icl::interval_set<int>>{
make_pair(boost::icl::discrete_interval<int>::closed(0,0), boost::icl::interval_set<int>{boost::icl::discrete_interval<int>::closed(3,3)})}));
cout << larger << endl;
cout << smaller << endl;
cout << boost::icl::contains(larger, smaller) << endl;
cout << ((smaller & larger) == smaller) << endl;
As given above, I'm trying to check if smaller is subset of larger. But I
get following output:
{([102,104]->{([0,0]->{[2,4]})})}
{([103,103]->{([0,0]->{[3,3]})})}
0
1
Any reason why boost::icl::contains(larger, smaller) doesn't work but
((smaller & larger) == smaller) works?
With this code:
size = 100;
uint64_t work;
row.get_value(3, work);
cout << "value was " << work << endl;
work += size;
cout << "value set to " << work << endl;
row.set_value(3, work);
row.get_value(3, work);
cout << "value now " << work << endl;
I expect this output:
value was 0
value set to 100
value now 100
but I get:
value was 0
value set to 100
value now 0
The updated value, 100, does display correctly in the tree view widget, I just cannot read it with get_value. What am I doing wrong?
Turns out the problem was the uint64_t; row[3] was defined (in Glade) as a guint, the work variable must match that type exactly or get_value will not work.
For example we have expression using rdivide in Matlab:
B = bsxfun(#rdivide, A, A(4,:));
How can we write equavalent expression for opencv?
Opencv has divide function, but seems it can't be used for matrix with different dimensions:
Mat t1= Mat::ones(2,3,CV_64FC1);
Mat t2= Mat::ones(1,3,CV_64FC1);
Mat dst;
divide(t1,t2,dst);
this don't work, so we need to replicate one row to matrix to match dimensions of t1 or use divide with 1 row in cycle.
My solution for opencv(A modified inplace):
for(int i=0;i<A.rows;++i)
{
divide(A.row(i),A.row(3),A.row(i));
}
Is there any simpler way?
You can use the repeat function of OpenCV to replicate a matrix.
The equivalent OpenCV code for the above mentioned MATLAB command is following:
cv::Mat B = A/cv::repeat(A.row(3),4,1);
In addition to #sgarizvi solution, you may find this wrapper to Matlab rdivide helpful:
#include <opencv2\opencv.hpp>
#include <iostream>
using namespace std;
using namespace cv;
Mat rdivide(const Mat& A, const Mat& B)
{
int nx = A.cols / B.cols;
int ny = A.rows / B.rows;
return A / repeat(B, ny, nx);
}
Mat rdivide(const Mat& A, double d)
{
return A / d;
}
int main()
{
Mat1f A = (Mat1f(3, 5) << 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15);
Mat B = rdivide(A, A.row(2)); // Divide by matrix, works also for cols: e.g. A.col(2)
Mat C = rdivide(A, 2); // Divide by scalar
cout << "A: " << endl << A << endl << endl;
cout << "B: " << endl << B << endl << endl;
cout << "C: " << endl << C << endl << endl;
return 0;
}
I want to have a goocanvas with a rect item which contains a text.
How to build a group with a rect to add the text as a child is known. But I want to know the size of the text to give the box ( rect ) enough space to display the complete text.
txt->property_width()
delivers 0! :-(
Glib::RefPtr<Goocanvas::Text> txt = Goocanvas::Text::create( "W123", 0, 0 );
Goocanvas::Bounds b = txt->get_bounds();
cout << b.get_x1() << endl;
cout << b.get_x2() << endl;
cout << b.get_y1() << endl;
cout << b.get_y2() << endl;
All values are zero!
Wow! If I change the code to:
Glib::RefPtr<Goocanvas::Text> txt = Goocanvas::Text::create( "W123", 0, 0 );
Goocanvas::Bounds b = txt->get_bounds();
GooCanvasBounds bounds;
goo_canvas_item_get_bounds (( GooCanvasItem*)txt->gobj(), &bounds);
cout << b.get_x1() << endl;
cout << b.get_x2() << endl;
cout << b.get_y1() << endl;
cout << b.get_y2() << endl;
The output is as expected! Attention: I look not at the bounds var! I still look for
'b'. Looks strange!
[Edit] I wrote a bug report: https://bugzilla.gnome.org/show_bug.cgi?id=721627
If your text is a GooCanvasText item, all you need is (in gtk+)
void goo_canvas_item_get_bounds (GooCanvasItem *item, GooCanvasBounds *bounds)
This translates to txt->get_bounds() in the gtkmm case. See goocanvas online devhelp for details.