I am using OpenCV 3.3.0 for c++ in visual studio. The problem is that when I use this code, it opens two windows with the same name that one of them contains the image and the another one is extra and empty. (Screenshot from the two windows )
cv::Mat im0 = cv::imread("C:\\Users\\MY_PC\\Desktop\\Image.bmp"); //read the image
cv::resize(im0, I, cv::Size(640, 480), 0, 0, CV_INTER_LINEAR); //Resize the image to 640x480
cv::namedWindow("HI");
cv::imshow("HI", I);
cv::waitKey(0);
The more important problem following this problem, is that the waitkey() function just works with the extra window and I should press the keys on that window to see the changes in the another window.
Furthermore the setMouseCallback has the same problem and mouse commands doesn't work on the window containing the image. Thanks a lot.
I have the same problem with the project type "Win32 Console Application".
If you use a "Empty Project" the problem disappears.
Try commenting out the line cv::namedWindow("HI");, does it work?
Unfortunately I can't help with other two.
I don't see any mistakes in the code you shared.
However, I'm usually writing like that:
const char* display_name = "Display";//Display name
namedWindow(display_name, WINDOW_AUTOSIZE);//Create Display
imshow(display_name, frame);//Display Image
waitKey(0);//Wait a key to be pressed
For x64: Removed opencv_world410.lib and just add opencv_world410d.lib from Additional Dependencies
(Common properties->Linker->Inputs->Additional Dependencies)
This link helps you to create config files for all of your projects:
https://docs.opencv.org/4.1.0/dd/d6e/tutorial_windows_visual_studio_opencv.html
Related
Hey guys I'm new in c++ wxwidgets programming.
I would like to know the easiest way to put an image into a button.
I tried :
button1 = new wxBitmapButton(side_panel, wxID_ANY, wxBitmap("image.png",wxBITMAP_TYPE_PNG), wxPoint(150,30), wxSize(30, 30),wxBORDER_NONE);
But I always get the same error:
If you expand the error dialog you see, you should see more information about the error, but my guess is that the image simply can't be found. You should check that the file image.png indeed exists in the current working directory of your program, i.e. the directory that you run it from, assuming you don't change it later.
You should also actually check for errors in your programs, even simple ones, i.e.
wxBitmap bmp("image.png");
if (!bmp.IsOk()) {
... handle the error somehow instead of blithely using an invalid bitmap ...
}
I'm pretty new to QT so maybe the answer to this question is simple. I'm making a terminal for our freescale car this year which is just a car that autonomously goes around a track. We're using line-scan cameras which give you a line of data 128 pixels long that you refine into either "white" or "black" (edges of the track) values. Last year the team made this terminal in QT which is used for modifying certain coefficients, stopping the motors, and all kinds of stuff through bluetooth. A text browser was used to display what the line cameras see.
Last year they combined the two cameras to create one image, but we're re configuring them this year, and they will produce two different images of the track. Naturally, I need to make two text browsers, one for each camera. Should simply be a case of copying code and changing some names right? Well apparently not.
I placed the text browser in the window using the design gui, named it, and made its settings match the other browser. I went into the code for the main window c file and adapted the code there. For every reference to the other browser I made sure one was an equal line of code for this one. When I was done and went to build the project, it told me that the browser wasn't a member of 'UI::MainWindow,' and pointed me to the ui_mainwindow.h file. So I went into that header and did the same thing in there. But when I built that, it gave me the error:
...\ui_mainwindow.h:1363: error: C2065: 'tb_camera_out_close' : undeclared identifier
How can that be? There doesn't appear to be any other reference to the other text browser and it works fine. How do I go about declaring this? Shouldn't the gui interface have done it for me when I placed it?
Here's the code from the main window's C file that is involved:
ui->tb_camera_out->setText(msgs_received[1] + '\n' + ui->tb_camera_out->toPlainText());
ui->tb_camera_out_close->setText(msgs_received[2] + '\n' + ui->tb_camera_out_close->toPlainText());
tb_camera_out is the old camera code(I am aware that the toplaintext() part won't work, I'm also attempting to get the text to scroll down instead of up like a track would if you were actually driving but I need to fix this major issue first)
Here's the code from the header that I added:
tb_camera_out = new QTextBrowser(centralWidget);
tb_camera_out->setObjectName(QStringLiteral("tb_camera_out"));
sizePolicy.setHeightForWidth(tb_camera_out->sizePolicy().hasHeightForWidth());
tb_camera_out->setSizePolicy(sizePolicy);
tb_camera_out->setMinimumSize(QSize(450, 0));
tb_camera_out->setMaximumSize(QSize(16777215, 16777215));
tb_camera_out->setStyleSheet(QLatin1String("background-color: rgb(0, 0, 0);\n" "color: rgb(31, 234, 0);"));
gridLayout_4->addWidget(tb_camera_out, 1, 0, 1, 5);
tb_camera_out_close = new QTextBrowser(centralWidget);
tb_camera_out_close->setObjectName(QStringLiteral("tb_camera_out_close"));
sizePolicy.setHeightForWidth(tb_camera_out_close->sizePolicy().hasHeightForWidth());
tb_camera_out_close->setSizePolicy(sizePolicy);
tb_camera_out_close->setMinimumSize(QSize(450, 0));
tb_camera_out_close->setMaximumSize(QSize(16777215, 16777215));
tb_camera_out_close->setStyleSheet(QLatin1String("background-color: rgb(0, 0, 0);\n" "color: rgb(31, 234, 0);"));
Again, the old code is at the top with the references to tb_camera_out. The stuff I added is associated with tb_camera_out_close.
I seem to have solved the problem. Strangely enough, to fix the errors that were showing up I not only had to edit that header, but I then had to delete it and run qmake to restore it. I then had to delete all the past builds to get the thing to even show up when I ran it.
I am trying to create a simple panel for Openbox in Arch Linux using c++, but I cannot figure out how to remove the title bar from a window.
I am creating the window with XCreateWindow(...), and that gives a window with the correct size, but it contains a title bar, and the window also opens in the top-left corner of the screen, no matter what offset coordinates I specify.
I read here that both of these problems are probably caused by the window manager (Openbox), which overrides the window attributes I specified in XCreateWindow(..., &window_attributes). This could be solved by adding window_attributes.override_redirect = True;, although this does not seem to do anything for me. When I try this I get the exact same window as before. (I did compile the file after this change.)
Also I read into the code of Tint2 (link), which is another panel for Openbox. They create a window using the following code:
XSetWindowAttributes att = { .colormap=server.colormap, .background_pixel=0, .border_pixel=0 };
p->main_win = XCreateWindow(server.dsp, server.root_win, p->posx, p->posy, p->area.width, p->area.height, 0, server.depth, InputOutput, server.visual, mask, &att);
I don't see an override_redirect anywhere in their code, so I'm not sure how they are removing the title bar.
As additional information, I thought it would be worth mentioning how I'm executing the script:
/* The c++ file is saved as 'panel.cpp' */
$ gcc panel.cpp -lX11 -o panel
$ ./panel
Also, I am running Arch Linux through VirtualBox with Windows 8 as host. I'm not sure if this changes anything, but it won't hurt to mention.
Since I found the solution, I figured I'd post the solution here if anyone else needs it.
As #JoachimPileborg mentioned, I needed to alter the Openbox settings in ~/.config/openbox/rc.xml. Inside the <applications> tag, I added the following code:
<application class="*">
<decor>no</decor>
<position force="no"></position>
</application>
The class="*" means that all applications will follow these rules, you could fill in the class name of the application instead. The <decor>no</decor> removes the title bar, and <position force="no"></position> ensures that my own script is able to handle the positioning. You could also add another <application> tag after this one to make exceptions to this rule.
Also, the window_attributes.override_redirect = True; is not needed anymore.
A more correct way is to use the Extended Window Manager Hints.
The idea is that you don't tell the window manager how to decorate or not your window, you just indicate the window type with _NET_WM_WINDOW_TYPE :
Atom window_type = XInternAtom(display, "_NET_WM_WINDOW_TYPE", False);
long value = XInternAtom(display, "_NET_WM_WINDOW_TYPE_DOCK", False);
XChangeProperty(display, your_window, window_type,
XA_ATOM, 32, PropModeReplace, (unsigned char *) &value,1 );
"Dock" is the type for panels and taskbar. Usually they are undecorated and appear on all desktops. As written on the documentation, previously the _MOTIF_WM_HINTS property was used to define the appearance and decorations of the window. Window managers still support it, but _NET_WM_WINDOW_TYPE is prefered as it describe the function and let the window manager (and user) decide on the appearance and behavior of that type of window.
Another interesting property for a panel is _NET_WM_STRUT_PARTIAL, to "reserve" space.
In three ways I tried to display an image within a Windows Forms Picturebox using Visual C++ 2010 Express. The files path I fetch with an openFileDialog. Here are the three attempts:
1.
pictureBox1->ImageLocation = openFileDialog1->FileName;
2.
Bitmap^ image1;
image1 = gcnew Bitmap(openFileDialog1->FileName, true);
pictureBox1->Image = image1;
3.
pictureBox1->Image = static_cast<Bitmap ^>(Image::FromFile(openFileDialog1->FileName));
Before the image is displayed finally an error- or info-message pops up. The message window has no name an shows BM6 (opening a Bimap-Image), nothing (opening a JPEG-Image) or ?PNG[] (opening a PNG-Image).
Question: How I can avoid those messages when displaying an image within a picturebox?
I solved my programming issue again on my own. The reason why a MessageBox showed those strange messages when an image was opened via openfiledialog was the code line MessageBox::Show(sr->ReadToEnd());. Probably a part of the Image-Header was displayed in the Box. I don't need it, so the solution is to delete or to comment out this line of code.
I am trying to use Unity, with a build script that creates my application by ultimately invoking BuildPipeline from the script. I am trying to figure out how to set up the icons, however. After calling PlayerSettings.SetIconsForTargetGroup and invoking BuildPipline.BuildPlayer, the appropriate icon does not show up for the executable file produced, nor display when the program is running.
I am currently using the following code.
Texture2D texture = AssetDatabase.LoadMainAssetAtPath(iconFile) as Texture2D;
int [] sizeList = PlayerSettings.GetIconSizesForTargetGroup(BuildTargetGroup.Standalone);
Texture2D[] iconList = new Texture2D[sizeList.Length];
for(int i=0;i<sizeList.Length;i++)
{
int iconSize = sizeList[i];
iconList[i] = (Texture2D)Instantiate(texture);
iconList[i].Resize(iconSize,iconSize,TextureFormat.ARGB32,false);
}
PlayerSettings.SetIconsForTargetGroup(BuildTargetGroup.Standalone,iconList);
What am I doing wrong?
Any assistance would be greatly appreciated. Thank you
Using BuildTargetGroup.Unknown works for me.
This also worked for me:
BuildTargetGroup.Unknown
But with an unwanted outcome: Unity does not override all icon sizes when setting one single Texture2D PNG, so Windows Explorer swaps the icon of my application EXE between the Unity default icon and my icon for different resolutions. Definitely, seems like a bug in Unity3D BuildPipeline.BuildPlayer(..) or SetIconsForTargetGroup(..).
If you're calling BuildPipline.BuildPlayer for OSX, then you need to specify full folder:
/MyPath/MyApp.app
i.e. don't leave out the ".app" part, because without that Unity builds the app, but doesn't include icons or splash image in resolution dialog.