I want to implement bag of words in opencv. after detector->detect(img, keypoint); detects keypoints, when i want to clean keypoints using keypoint.clear(); or when the function wants to return the following error will be appeared.
"Unhandled exception at 0x011f45bb in BOW.exe: 0xC0000005: Access violation reading location 0x42ebe098."
and also detected keypoints have bizarre points coordinates like cv::Point_ pt{x=-1.5883997e+038y=-1.5883997e+038 }
Part of the code
Ptr<DescriptorMatcher> matcher = DescriptorMatcher::create("FlannBased");
Ptr<DescriptorExtractor> extractor = new SurfDescriptorExtractor();
Ptr<FeatureDetector> detector = new SurfFeatureDetector(2000);
void extractTrainingVocabulary() {
IplImage *img;
int i,j;
CvSeq *imageKeypoints = 0;
for(j=1;j<=60;j++)
for(i=1;i<=60;i++){
sprintf( ch,"%d%s%d%s",j," (",i,").jpg");
const char* imageName = ch;
Mat img = imread(ch);
vector<KeyPoint> keypoint;
detector->detect(img, keypoint);
Mat features;
extractor->compute(img, keypoint, features);
bowTrainer.add(features);
keypoint.clear();//problem
}
return;
}
I noticed something about your code, on extractTrainingVocabulary() you declare IplImage* img; and inside the loop you declare another variable with the same name (but different type): Mat img = imread(ch);.
Even though that might not be the problem, it's certainly not good practice. I would fix that immediately and update the code on your question.
Related
I know this "looks" like a duplicate question, but -- it's not answered in any definitive way that I can see.
In C++, OpenCv operation to init a Mat can be done like so:
float[,] camera = new float[,] {
{ 857.483f, 0.0f, 968.06f },
{ 0.0f, 876.72f, 556.37f },
{ 0.0f, 0.0f, 1.0f }
};
A = Mat(3,3, CV_32FC1, &camera);
Note that the array is 2D.
Now, I want to accomplish the same thing in EMGU, //using managed C#; so Mat is constructed as:
Mat _cameraMatrix = new Mat(3, 3, DepthType.Cv64F, 1);
// and a similar 1D Mat:
Mat _distCoeffs = new Mat(8, 1, DepthType.Cv64F, 1);
// to be passed to the method:
CvInvoke.Undistort(bkgimage, ndimage, _cameraMatrix, _distCoeffs);
Unlike cv::Mat, EMGU Mat's constructor doesn't seem to have an obvious way to get that 2D "camera" data into it. Anyone have any success getting this kind of code to work?
Funnily enough, I came across this question while trying to solve exactly the same problem - camera calibration.
There's an easier but much less elegant method to solve this:
Image<Gray, float> camImage = new Image<Gray, float>(new float[,,]
{
{intrinsics.fx}, {0}, {intrinsics.ppx}},
{ {0}, {intrinsics.fy}, {intrinsics.ppy}},
{ {0}, {0}, {1} }
});
var camMatrix = camImage.Mat;
Image<Gray, float> distImage = new Image<Gray, float>(new float[,,] {{{0},{0},{0},{0}}});
var distCoeffs = distImage.Mat;
Not sure it's the best answer, but it's an answer.
According to the EmguCV documentation, there is a constructor that takes an IntPr as 'data':
see http://www.emgu.com/wiki/files/3.4.3/document/html/c8424736-2d44-c5d6-212f-31dedfe6fb95.htm
The problem is how to get that pointer. Here, maybe this post will help you:
How to get IntPtr from byte[] in C#
Hopefully, this will get you at least started.
I am trying to save an OpenCV image to the hard drive.
Here is what I tried:
public void SaveImage (Mat mat) {
Mat mIntermediateMat = new Mat();
Imgproc.cvtColor(mRgba, mIntermediateMat, Imgproc.COLOR_RGBA2BGR, 3);
File path =
Environment.getExternalStoragePublicDirectory(
Environment.DIRECTORY_PICTURES);
String filename = "barry.png";
File file = new File(path, filename);
Boolean bool = null;
filename = file.toString();
bool = Highgui.imwrite(filename, mIntermediateMat);
if (bool == true)
Log.d(TAG, "SUCCESS writing image to external storage");
else
Log.d(TAG, "Fail writing image to external storage");
}
}
Can any one show how to save that image with OpenCV 2.4.3?
Your question is a bit confusing, as your question is concerning OpenCV on the desktop, but your code is for Android, and you ask about IplImage, but your posted code is using C++ and Mat. Assuming you're on the desktop using C++, you can do something along the lines of:
cv::Mat image;
std::string image_path;
//load/generate your image and set your output file path/name
//...
//write your Mat to disk as an image
cv::imwrite(image_path, image);
...Or for a more complete example:
void SaveImage(cv::Mat mat)
{
cv::Mat img;
cv::cvtColor(...); //not sure where the variables in your example come from
std::string store_path("..."); //put your output path here
bool write_success = cv::imwrite(store_path, img);
//do your logging...
}
The image format is chosen based on the supplied filename, e.g. if your store_path string was "output_image.png", then imwrite would save it was a PNG image. You can see the list of valid extensions at the OpenCV docs.
One caveat to be aware of when writing images to disk with OpenCV is that the scaling will differ depending on the Mat type; that is, for floats the images are expected to be within the range [0, 1], while for say, unsigned chars they'll be from [0, 256).
For IplImages, I'd advise just switching to use Mat, as the old C-interface is deprecated. You can convert an IplImage to a Mat via cvarrToMat then use the Mat, e.g.
IplImage* oldC0 = cvCreateImage(cvSize(320,240),16,1);
Mat newC = cvarrToMat(oldC0);
//now can use cv::imwrite with newC
alternately, you can convert an IplImage to a Mat just with
Mat newC(oldC0); //where newC is a Mat and oldC0 is your IplImage
Also I just noticed this tutorial at the OpenCV website, which gives you a walk-though on loading and saving images in a (desktop) environment.
I'm trying to use facial recognition with Android . All the loads are ok , but the haarcascade_frontalface_alt2.xml file wich i don't know how to load it using JavaCV.
This is the code i have:
public static void detectFacialFeatures()
{
// The cascade definition to be used for detection.
// This will redirect the OpenCV errors to the Java console to give you
// feedback about any problems that may occur.
new JavaCvErrorCallback();
// Load the original image.
// We need a grayscale image in order to do the recognition, so we
// create a new image of the same size as the original one.
IplImage grayImage = IplImage.create(iplImage.width(),iplImage.height(), IPL_DEPTH_8U, 1);
// We convert the original image to grayscale.
cvCvtColor(iplImage, grayImage, CV_BGR2GRAY);
CvMemStorage storage = CvMemStorage.create();
// We instantiate a classifier cascade to be used for detection, using the cascade definition.
CvHaarClassifierCascade cascade = new CvHaarClassifierCascade(cvLoad("./haarcascade_frontalface_alt2.xml"));
// We detect the faces.
CvSeq faces = cvHaarDetectObjects(grayImage, cascade, storage, 1.1, 1, 0);
Log.d("CARAS","Hay "+faces.total()+" caras ahora mismo");
}
The problem is here
CvHaarClassifierCascade(cvLoad("./haarcascade_frontalface_alt2.xml"));
I have tried putting the xml file it into the /assets folder , but i have no idea of how must i load it. It's always giving me the next error:
03-26 17:31:25.385: E/cv::error()(14787): OpenCV Error: Null pointer (Invalid classifier cascade) in CvSeq* cvHaarDetectObjectsForROC(const CvArr*, CvHaarClassifierCascade*, CvMemStorage*, std::vector<int>&, std::vector<double>&, double, int, int, CvSize, CvSize, bool), file /home/saudet/projects/cppjars/OpenCV-2.4.4/modules/objdetect/src/haar.cpp, line 1514
...
looking more near at the error it points to this code line:
CvSeq faces = cvHaarDetectObjects(grayImage, cascade, storage, 1.1, 1,
0);
That's why i'm pretty sure that the problem comes from the haarcascade_frontalface_alt2.xml load.
Thanks for your help.
P.D: I want to include the cascade into the apk not in sdcard .
If your cascade is in SD card you can use:
CascadeClassifier cascade = new CascadeClassifier(Environment.getExternalStorageDirectory().getAbsolutePath() + "/cascade.xml");
Environment.getExternalStorageDirectory().getAbsolutePath() give you right path to SD card and next - is address to your file in your SD.
You can pack your file in apk and then copy it to external location so it is accessible by OpenCV functions.
try {
File learnedInputFile = new File(Environment.getExternalStorageDirectory().getPath() + "/learnedData.xml");
if (!learnedInputFile.exists()) {
InputStream learnedDataInputStream = assetManager.open("learnedData.xml");
FileOutputStream learnedDataOutputStream = new FileOutputStream(learnedInputFile);
// copy file from asset folder to external location, i.e. sdcard
byte[] buffer = new byte[300];
int n = 0;
while (-1 != (n = learnedDataInputStream.read(buffer))) {
learnedDataOutputStream.write(buffer, 0, n);
}
}
classifier.load(learnedInputFile.getAbsolutePath());
} catch (IOException exception) {
// there are no learned data, train ml algorithm or show error, etc.
}
I have the following code:
IplImage* f( IplImage* src )
{
// Must have dimensions of output image
IplImage* cropped = cvCreateImage( cvSize(1280,500), src->depth, src->nChannels );
// Say what the source region is
cvSetImageROI( src, cvRect( 0,0, 1280,500 ) );
// Do the copy
cvCopy( src, cropped );
cvResetImageROI( src );
return cropped;
}
void testApp::setup(){
img.loadImage("test.jpg");
finder.setup("haarcascade_frontalface_default.xml");
finder.findHaarObjects(img);
}
//--------------------------------------------------------------
void testApp::update(){
}
//--------------------------------------------------------------
ofRectangle cur;
void testApp::draw(){
img = f(img);
img.draw(0, 0);
ofNoFill();
for(int i = 0; i < finder.blobs.size(); i++) {
cur = finder.blobs[i].boundingRect;
ofRect(cur.x-20, cur.y-20, cur.width+50, cur.height+50);
}
}
It produces an error. I think it's because I don't convert IplImage to ofImage. Can someone please tell me how to do it?
I would imagine that your instance of img, that your passing to f(), is an instance of ofxCVImage or similar. As far as I know ofx store a protected variable of iplimage, as such you can't cast ofxCVImage to an iplimage.
You might try
img = f(img.getCvImage()) // ofxCvImage::getCvImage() returns IplImage
I am experimenting with JavaCV (OpenCV) and I am confused with the following behavior.
My programm simply:
Grab an image
Create a grayscale version of the image (leaving the original image untouched)
Threshold the grayscale image
Find contours in the grayscale image (cloning the image since cvFindContours modifies the image and we want to display it as is)
Draw contours on the original color image
The problem is that unless I clone the grabbedImage (see commented line), the grayscale image is modified and contours are drawn on it. Also, it is as if multiple contours are drawn on the grabbedImage.
I also tried to add a sleep to the loop and it fixes the problem. Could it be that I get the same (modified) grabbedImage multiple time? I checked the java reference and it's different but could it be the same buffer?
Any idea?
Thank you
package com.mdarveau.opencvtest;
import static com.googlecode.javacv.cpp.opencv_core.*;
import static com.googlecode.javacv.cpp.opencv_imgproc.*;
import com.googlecode.javacpp.Loader;
import com.googlecode.javacv.CanvasFrame;
import com.googlecode.javacv.FrameGrabber;
import com.googlecode.javacv.cpp.opencv_core.CvContour;
import com.googlecode.javacv.cpp.opencv_core.CvMemStorage;
import com.googlecode.javacv.cpp.opencv_core.CvScalar;
import com.googlecode.javacv.cpp.opencv_core.CvSeq;
import com.googlecode.javacv.cpp.opencv_core.IplImage;
import com.googlecode.javacv.cpp.opencv_objdetect;
public class Demo {
public static void main( String[] args ) throws Exception {
// Preload the opencv_objdetect module to work around a known bug.
Loader.load( opencv_objdetect.class );
FrameGrabber grabber = FrameGrabber.createDefault( 1 );
grabber.start();
IplImage grabbedImage = grabber.grab();
int width = grabbedImage.width();
int height = grabbedImage.height();
IplImage grayImage = IplImage.create( width, height, IPL_DEPTH_8U, 1 );
CvMemStorage storage = CvMemStorage.create();
CanvasFrame filterProbe = new CanvasFrame( "Filtered", CanvasFrame.getDefaultGamma() / grabber.getGamma() );
CanvasFrame enhancedProbe = new CanvasFrame( "Enhanced", CanvasFrame.getDefaultGamma() / grabber.getGamma() );
while ( filterProbe.isVisible() && enhancedProbe.isVisible() && (grabbedImage = grabber.grab()) != null ) {
cvClearMemStorage( storage );
// Convert to grayscale image...
cvCvtColor( grabbedImage, grayImage, CV_BGR2GRAY );
// UNCOMMENT FIXES THE PROBLEM grabbedImage = grabbedImage.clone();
// Let's find some contours! but first some thresholding...
cvThreshold( grayImage, grayImage, 128, 255, CV_THRESH_BINARY );
// To check if an output argument is null we may call either isNull() or equals(null).
CvSeq contour = new CvSeq( null );
// cvFindContours modifies the image so clone it first since we want to keep the grayscale version
cvFindContours( grayImage.clone(), storage, contour, Loader.sizeof( CvContour.class ), CV_RETR_LIST, CV_CHAIN_APPROX_SIMPLE );
while ( contour != null && !contour.isNull() ) {
if ( contour.elem_size() > 0 ) {
CvSeq points = cvApproxPoly( contour, Loader.sizeof( CvContour.class ), storage, CV_POLY_APPROX_DP, cvContourPerimeter( contour ) * 0.02, 0 );
cvDrawContours( grabbedImage, points, CvScalar.BLUE, CvScalar.BLUE, -1, 1 /*CV_FILLED*/, CV_AA );
}
contour = contour.h_next();
}
filterProbe.showImage( grayImage );
enhancedProbe.showImage( grabbedImage );
}
filterProbe.dispose();
enhancedProbe.dispose();
grabber.stop();
}
}
This is the expected behavior of cvFindContours():
The function modifies the image while extracting the contours.