I am using a Haar cascade classifier trained from the MIT cars dataset to detect vehicles in OpenCV (trained using the utilities provided with OpenCV). This works reasonably well when compiled in Debug mode, but when compiled in Release mode the cascade doesn't make any detections at all. Running the following code on the test image below gives a detection in debug mode but nothing in release mode (this behaviour continues through all images in my data sequence).
Can you suggest why this occurs and, more importantly, what I can do to obtain detections when running in Release mode?
Cascade File
Code
cv::Mat testImage = cv::imread("testImage.png",0);
cv::equalizeHist(testImage, testImage);
cv::CascadeClassifier vehicleCascade;
vehicleCascade.load("cars3.xml");
// Detect vehicles
std::vector<cv::Rect> cars;
vehicleCascade.detectMultiScale(
testImage, // Input image
cars, // Output bounding boxes
1.1, // scale factor - how much image size is reduced at each scale
5, // min neighbours - how many neighbours required to maintain rect
0|CV_HAAR_SCALE_IMAGE, // Not used
cv::Size(30,30), // Min poss object size
cv::Size() // Max poss object size
);
std::cout << "Found " << cars.size() << " objects.\n";
for (int i=0; i<cars.size(); ++i)
cv::rectangle(testImage, cars.at(i), CV_RGB(255,0,0), 3);
cv::namedWindow("Haar cascade");
cv::imshow("Haar cascade", testImage);
cv::waitKey(0);
cv::imwrite("output.png", testImage);
TestImage
It's strange, but for me you code is working fine in both modes. I'm using Visual Studio 2010 on Windows 7 32bit. Here is my project - https://www.dropbox.com/s/5kubn5tlu7k6ziy/opencvhw.rar, so you can check executables(Release and Debug directories). If you are using visual studio and want to build it on your own you will have to change paths to library and include directories for both modes(include directories path is the same, library directories path is different). (project -> ... properties -> configuration properties -> vc++ directories)
Generally i would recommend to check really carefully project configuration. It's really easy to make small mistake which can cause very weird behavior. Sometimes the best option is to configure everything from the scratch once again.
OpenCV have very strange bugs - it's nothing new :) For example i can't use some codecs while debugging - if i run program in debug mode, but without debugging everything is fine, but if i try to debug - each frame readed from file is empty(but its size is correct).
It's possible that you have just encountered something a bit similar. Try to convert image to different format(i think that bmp is the best choice for test - it's should always work without any additional libraries).
Also note that you have uploaded jpg file so i had to rename and convert it on my own - maybe during conversions something in this file have changed so we are not testing your code on exactly the same files - upload you png file on dropbox so i can test it.
See this post:
This type of error could be caused by linking to the debug library of opencv in release mode.
Related
I need to check elements of some cv::Mats while debugging my OpenCV C++ code.
Tried adding mymat.at<double>(0, 0) to watch but there are two problems:
Sometimes it doesn't work: Couldn't find method cv::Mat::at<double>
Even if it works, it is very hard to check elements one by one.
Currently, I write Mat contents to a file using cv::FileStorage but it is not a good solution also.
Is there any VSCode extension like Image Watch to show value of Mat elements while debugging?
I am starting to learn object detection in OpenCV 3.4.2 (.Net C++ 2017).
I am very interested in detecting strawberries in pictures (at the moment I am really interested in detecting just strawberries). I know OpenCV has some pre-trained Haar cascade files in OpenCV directory, but there are not .xml files for strawberries (there are for body parts instead).
So I decided to search on Google, to try to find trained strawberry Haar cascade .xml files. I found this .xml file XML strawberry file but I get error -49 when I try to execute the program. I have executed correctly the program using OpenCV files, but I cannot execute correctly when I try with GitHub XML file.
I have found this thread here on StackOverflow StackOverFlow thread about GitHub XML files in OpenCV and a user claims that it's not possible to use GitHub XML files into OpenCV.
My question is about if there is a way to use the XML GitHub file I have posted in this thread in OpenCV or I need to train my own XML file? I would like to use GitHub file.
Edit(1)
I have found this link strawberry detection in OpenCV where, if you look at the source code, it seems that the same strawberry_classifier.xml is being used. I don't know if the name of the file is just a coincidence (Github filename and the filename shown in the source code of the 3rd link are exactly the same). At least it seems that the programmer (from the 3rd link) has obtained some results while using the (apparently) same .xml file that I want to use. But I don't know how to use that strawberry_classifier.xml file.
Python dev here,
I'm late, but in case anyone still wants to see an answer:
The classifier from GitHub works perfectly fine as shown in this Python code (Sorry, I didn't do it in C++, but I think it won't be much different)
The script uses your webcam as the image source. You can show the webcam some images of strawberries, and it will recognize it:
import cv2 #import library
#define Haar Cascade Classifier
Strawberry_Classifier = cv2.CascadeClassifier(r"C:\Users\Strawberry.xml")
VideoCapture = cv2.VideoCapture(0) #capture video from camera
#set video size
VideoCapture.set(3, 540)
VideoCapture.set(4, 360)
while True:
#Connect video and convert
Connection_Success, Video = VideoCapture.read() #returns a bool and video array in one tuple (sucess, video array)
RGB_video = cv2.cvtColor(Video, cv2.COLOR_BGR2RGB) #converts to suitable format
Detect_Strawberry = Strawberry_Classifier.detectMultiScale(RGB_video, 1.3, 13) #MODIFY THIS FOR LESS/MORE DETECTION ACCURACY
#Detect Strawberry
for(x,y,w,h) in Detect_Strawberry: #x,y width, height
cv2.rectangle(Video, (x, y), (x + w, y + h), (0, 255, 0), 3) #Put a rectangle around Strawberry
cv2.imshow("Window", Video) #show video
#quit if q is pressed, quit
QuitKey = cv2.waitKey(30)
if QuitKey == ord("q"):
VideoCapture.release()
cv2.destroyAllWindows()
A very simple question...why am I getting a read access violation error with this code?
cv::Mat laserSpeckle = Mat::zeros(100,100,CV_8UC1);
imwrite( "C://testimage.jpg", laserSpeckle );
When i attach a debugger and look into it further, it throws the exception at this snippet in grfmt.cpp.
if( params[i] == CV_IMWRITE_JPEG_QUALITY )
{
quality = params[i+1];
quality = MIN(MAX(quality, 0), 100);
}
It occurs with .png and .tiff too. Im an OpenCV newbie, my apologies if this is something really simple. I am using Qt for what its worth.
Do you build OpenCV yourself? If yes, make sure that the option WITH_JPEG is enabled when you configure your build files:
cmake ... -DWITH_JPEG=ON ...
If you want to save image with alpha channel you should use png format. It is described here
It should work with bmp format:
cv::Mat laserSpeckle = cv::Mat::zeros(100,100,CV_8UC1);
cv::imwrite( "C://testimage.bmp", laserSpeckle );
Your code also works on my computer. However, it seems that on some systems it works only for bmp images. I saw similar issues reported here and here.
The problem is with the debugger version (x64) if you build the code using the release version (x64) it works fine for me.
In my case c++ Code Generation settings were wrong
Should have been Multithreaded DEBUG dll MD
This is driving me up the wall..
I've got a very simple SDL2 program.
It has a array of 3 SDL_Texture pointers.
These textures are filled as follows:
SDL_Texture *myarray[15];
SDL_Surface *surface;
for(int i=0;i<3;i++)
{
char filename[] = "X.bmp";
filename[0] = i + '0';
surface = SDL_LoadBMP(filename);
myarray[i] = SDL_CreateTextureFromSurface(myrenderer,surface);
SDL_FreeSurface(surface);
}
This works, no errors.
In the main loop (which is just a standard event loop waiting for SDL_QUIT, keystrokes and a user-event which a SDL_Timer puts in the event queue every second) I just do (for the timer triggered event):
idx = (idx+1) % 3; // idx is global var initially 0.
SDL_RenderClear(myrenderer);
SDL_RenderCopy(myrenderer, myarray[idx], NULL, NULL);
SDL_RendererPresent(myrenderer);
This works fine for 0.bmp and 1.bmp, but the 3rd image (2.bmp) simply shows as a black field.
This is structural.
If I alternate the first 2 images they are both fine.
If I alternate the 2nd and 3rd image the 3rd image doesn't show.
If I use more than 3 images then 3 and upwards show as black.
Loading order doesn't matter. It starts going wrong with the 3rd image loaded from disk.
All images are properly formatted BMP's.
I even saved 2.bmp back to disk under a different name by using SDL_SaveBMP() after it was loaded to make sure it got loaded in memory OK. The new file is bit for bit identical to the original.
This program, without modifications and the same bmp files, works fine on OSX (XCode5) and Windows (VC++ 2012 Express).
The problem only shows on the Raspberry PI.
I have placed explicit error checks on every call that can leave a result/error-code (not shown in the samples above for brevity) but all of them show "no error".
I have used the latest stable source set of www.libsdl.org and compiled as instructed (configure, make, make install, etc.).
Anybody got any idea what could be going on ?
P.S.
Keyboard input doesn't seem to work either on my PI, but I haven't delved into that yet.
Answering myself as I finally figured it out myself...
I finally went back to the README-raspberrypi.txt that came with the SDL2 sources.
I didn't read it carefully enough the first time around...
Problem 1: I'am running on a FULL-HD display. The PI's default GPU memory is 64MB which is not enough for large displays and double-buffering. As suggested in the README I increased this to 128MB and this solved the black image problem.
Problem 2: Text input wasn't working because my user-account was not in the input group. I had added the default "pi" account to the input group initially, but when I later started using another account I forgot to add that user to the group.
In short: Caught by my own (too) quick skimming of the documentation.
I am new to c++ programming , today i was trying to save an image using CImg .
CImg is C++ Template Image Processing Library .
The basic code i wrote is(Please forgive any syntax erros , as copied part of my codes) :
#include "CImg.h"// Include CImg library header.
#include <iostream>
using namespace cimg_library;
using namespace std;
const int screen_size = 800;
//-------------------------------------------------------------------------------
// Main procedure
//-------------------------------------------------------------------------------
int main()
{
CImg<unsigned char> img(screen_size,screen_size,1,3,20);
CImgDisplay disp(img, "CImg Tutorial");
//Some drawing using img.draw_circle( 10, 10, 60, RED);
img.save("result.jpg"); // save the image
return 0;
}
But I cannot run my program as it says :
Invalid Parameter - 100%
'gm.exe' is not recognized as an internal or external command,
operable program or batch file.
This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.
[CImg] *** CImgIOException *** [instance(800,800,1,3,02150020,non-shared)] CImg<unsigned char>::save_other() : Failed to save file 'result.jpg'. Format is not natively supported, and no external commands succeeded.
terminate called after throwing an instance of 'cimg_library::CImgIOException'
what(): [instance(800,800,1,3,02150020,non-shared)] CImg<unsigned char>::save_other() : Failed to save file 'result.jpg'. Format is not natively supported, and no external commands succeeded.
Though i can see the image , I cannot save it.
After googling a bit i found people saying to install ImageMagick , i have installed it but no help .
Some of the Forum says to compile against libjpeg, libpng, libmagick++. But i don't know how to compile against those libraries.
I am using Eclipse CDT plugin to write C++ project .
Please help me .
I had the same error, and installing of GraphicsMagick (not ImageMagick) helped me.
I've downloaded and installed GraphicsMagick-1.3.26-Q8-win64-dll.exe from ftp://ftp.graphicsmagick.org/pub/GraphicsMagick/windows/. You may choose another one, if you need:
Note that the QuantumDepth=8 version (Q8) which provides industry
standard 24/32 bit pixels consumes half the memory and about 30% less
CPU than the QuantumDepth=16 version (Q16) which provides 48/64 bit
pixels for high-resolution color. A Q8 version is fine for processing
typical photos intended for viewing on a computer screen. If you are
dealing with film, scientific, or medical images, use ICC color
profiles, or deal with images that have limited contrast, then the Q16
version is recommended.
Important: during installation, don't remove checkbox "Update executable search path", which updates environment variable %PATH%, making gm.exe available from any place.
In my case, it was also required to install Ghostscript - which is highly recommended to install by GraphicsMagick. There is a link to x64 Ghostscript: https://sourceforge.net/projects/ghostscript/files/GPL%20Ghostscript/9.09/gs909w64.exe/download (I've put it here, because links from the GraphicMagick websites leads you to 32-bit only).
After that, it worked fine for me.
For some image formats (as .jpg, .png, .tif and basically all formats that require data compression), CImg will try to use an external tool to save them (such as convert from ImageMagick or gm from GraphicsMagick).
If you don't have any installed, then you won't be able to save .jpg files without having to link your code with the libjpeg library, to get a native support for JPEG read/write (then, you'll need to #define cimg_use_jpeg before #include "CImg.h", to tell the library you want to use the libjpeg features).
If you want to keep things simpler, I'd recommend to save your image using another (non-compressed) image format, as .bmp or .ppm.
These formats are handled natively by CImg and do not require to link with external libraries.
I know this question is old, but I kept getting the same error on one project and not on another and this is the only thing on Google.
To get rid of it, you must do 2 things:
Install dynamic ImageMagick libraries for your appropriate OS and architecture(32/64). Link
I was using VisualStudio, and the character set must be set to "Unicode". The error would appear again when I reverted back to Multi-Byte character set. I guess this has something to do with the way CImg handles strings and miscompares them.