Geppertm's profile - activity

You´ve got to use uEye-Lib to grab an Image. After that you can create a Mat with the width,highth and deepth of the grabed Image, then copy the Data from the Image to the Mat.

There are several models from econ systems which are uvc compliant. The 5-MP USB Cam gives you the full control using UVC. The USB 3.0 Models bring there own library to control them. http://www.e-consystems.com/default-de.asp

Great news, pleace keep me uptodate.

Hi, I´ve made a 7z File with the changes, you can find it here: http://www.altdorfer-jungs.de/2014/opencv.7z

the main changes are in cap_libv4l.cpp.

I think the best way would be to make a Function which set´s this value in v4l2. So the lib would be compatible to older source which also has such a Limit.

Unfortunatly, i don´t know where to upload my changes or add a bug for v4l.

It´s a "Bug" of video4Linux not openCV.

Hi, Yes I´m doing this. I´m using a 8MP USB3 Cam with Ubuntu 12.04.

You have to Edit the source of lib v4l. The problem ist the call of v4l2_read_and_convert(,..,..., int maxsize). There is a defined max size (16777216), if you change this to > 23970816 it will work.

I´m sorry that i actually couldn´t give you the File where to edit the value.

C/C++ Ubuntu:

system("shutdown 0");

C/C++ Windows

system("shutdown -f");

The Problem is that your are initializing RNG always with the same value. Try this:

#include <time.h>
.
.

void initializeAnts( SingleAnt *ants, Mat *sourceImage )
{
    uint64 initValue = time(0);
    RNG rng( initValue );
    int imgWidth = sourceImage->cols;
    int imgHight = sourceImage->rows;

    for( int index = 0; index < ANTSNUMBER; index++ ) {
        ants[ index ].col = rng.uniform( 0, imgWidth );
        ants[ index ].row = rng.uniform( 0, imgHight );
    }
}

This should give you different random values.

I´ve found out that this feature isn´t implemented for Linux. I´ve implemented the features. I´m going to commit this features as soon as possible.

I think you have to load the tif file into a char(or byte) array.(buf) Then use Mat imdecode(InputArray buf, int flags) Then you will get the size of the first image out of the returnend Mat. With the size of the first image you can delete it from the buffer and you will get the second image.

http://www.awaresystems.be/imaging/tiff/faq.html

I think the simples way ist to convert the Image to HSV, there you can easily change the color from red to yellow. After that you can convert it back to RGB(or BGR).

Best,

OpenCV is using BGR instead of RGB.

Yes it should do the same but I had the same Problem using VS2010 with Win 7 64-Bit using OpenCV 2.4.4. Using grab + retrieve instead of read solved it.

Use

cap.grab();
cap.retrieve(&frame, 0);

instead of

cap.read(frame);

Best

Hi,

Yes it is possible. I´m doing it with Data Matrix Codes. I´m creating regions using a threshold. Then i´m checking the anisometry and the size of the regions. The decoding is done by zxing.

Best,

Got the same problem using XUbuntu 12.04 + OpenCV 2.4.3.

Use this shell script it will do anything for you. Run it as superuser. [Sorry couldn´t attache it]

[istall_opencv.sh]

arch=$(uname -m)
if [ "$arch" == "i686" -o "$arch" == "i386" -o "$arch" == "i486" -o "$arch" == "i586" ]; then
flag=1
else
flag=0
fi
echo "Installing OpenCV 2.4.3"
mkdir OpenCV
cd OpenCV
echo "Removing any pre-installed ffmpeg and x264"
sudo apt-get remove ffmpeg x264 libx264-dev
echo "Installing Dependenices"
sudo apt-get install build-essential checkinstall cmake pkg-config yasm
sudo apt-get install libtiff4-dev libjpeg-dev libjasper-dev
sudo apt-get install libavcodec-dev libavformat-dev libswscale-dev libdc1394-22-dev libxine-dev libgstreamer0.10-dev libgstreamer-plugins-base0.10-dev libv4l-dev
sudo apt-get install python-dev python-numpy
sudo apt-get install libtbb-dev
sudo apt-get install libqt4-dev libgtk2.0-dev
echo "Downloading x264"
wget ftp://ftp.videolan.org/pub/videolan/x264/snapshots/x264-snapshot-20121126-2245-stable.tar.bz2
tar -xvf x264-snapshot-20121126-2245-stable.tar.bz2
cd x264-snapshot-20121126-2245-stable/
echo "Installing x264"
if [ $flag -eq 1 ]; then
./configure --enable-static
else
./configure --enable-shared --enable-pic
fi
make
sudo make install
cd ..
echo "Downloading ffmpeg"
wget http://ffmpeg.org/releases/ffmpeg-0.11.2.tar.bz2
echo "Installing ffmpeg"
tar -xvf ffmpeg-0.11.2.tar.bz2
cd ffmpeg-0.11.2/
if [ $flag -eq 1 ]; then
./configure --enable-gpl --enable-libfaac --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libtheora --enable-libvorbis --enable-libx264 --enable-libxvid --enable-nonfree --enable-postproc --enable-version3 --enable-x11grab
else
./configure --enable-gpl --enable-libfaac --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libtheora --enable-libvorbis --enable-libx264 --enable-libxvid --enable-nonfree --enable-postproc --enable-version3 --enable-x11grab --enable-shared
fi
make
sudo make install
cd ..
echo "Downloading v4l"
wget http://www.linuxtv.org/downloads/v4l-utils/v4l-utils-0.9.3.tar.bz2
echo "Installing v4l"
tar -xvf v4l-utils-0.9.3.tar.bz2
cd v4l-utils-0.9.3/
make
sudo make install
cd ..
echo "Downloading OpenCV 2.4.3"
wget -O OpenCV-2.4.3.tar.bz2 http://sourceforge.net/projects/opencvlibrary/files/opencv-unix/2.4.3/OpenCV-2.4.3.tar.bz2
echo "Installing OpenCV 2.4.3"
tar -xvf OpenCV-2.4.3.tar.bz2
cd OpenCV-2.4.3
mkdir build
cd build
cmake -D CMAKE_BUILD_TYPE=RELEASE ..
make
sudo make install
sudo echo “/usr/local/lib” >> /etc/ld.so.conf
sudo ldconfig
echo "OpenCV 2.4.3 ready to be used"

Hi, I would first use some kind of threshold on the image. So you have all ways white and all walls black.

Then I would create a Integer-array with the map of the Labyrinth.

Now you can easy use the algorythm first turn left, walk forward till there´s a Wall then turn right,walk forward till wall then turn right........ This will work for the most Labyrinths.

or

You search all impasse [if way = 0 and wall = 1 , all the fields that border on three ones] and fill them till the next crossing. That will give you all possible ways.[work´s everytime]

You can also use: Maze Algorythm from Gaston Tarry [work´s everytime]

Hey, the problem is that CV_CAP_PROP_FPS isn´t implemented yet. If you look at highgui/src/cap_libv4l2.cpp you´ll find this switch case in function icvGetPropertyCAM_V4L ...

  switch (property_id) {
case CV_CAP_PROP_FRAME_WIDTH:
case CV_CAP_PROP_FRAME_HEIGHT:
  CLEAR (capture->form);
  capture->form.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
  if (-1 == xioctl (capture->deviceHandle, VIDIOC_G_FMT, &capture->form)) {
      /* display an error message, and return an error code */
      perror ("VIDIOC_G_FMT");
    if (v4l1_ioctl (capture->deviceHandle, VIDIOCGWIN, &capture->captureWindow) < 0) {
      fprintf (stderr, "HIGHGUI ERROR: V4L:cap_libv4l: Unable to determine size of incoming image\n");
      icvCloseCAM_V4L(capture);
      return -1;
    } else {
      int retval = (property_id == CV_CAP_PROP_FRAME_WIDTH)?capture->captureWindow.width:capture->captureWindow.height;
      return retval / 0xFFFF;
    }
  }
  return (property_id == CV_CAP_PROP_FRAME_WIDTH)?capture->form.fmt.pix.width:capture->form.fmt.pix.height;
case CV_CAP_PROP_BRIGHTNESS:
  sprintf(name, "Brightness");
  capture->control.id = V4L2_CID_BRIGHTNESS;
  break;
case CV_CAP_PROP_CONTRAST:
  sprintf(name, "Contrast");
  capture->control.id = V4L2_CID_CONTRAST;
  break;
case CV_CAP_PROP_SATURATION:
  sprintf(name, "Saturation");
  capture->control.id = V4L2_CID_SATURATION;
  break;
case CV_CAP_PROP_HUE:
  sprintf(name, "Hue");
  capture->control.id = V4L2_CID_HUE;
  break;
case CV_CAP_PROP_GAIN:
  sprintf(name, "Gain");
  capture->control.id = V4L2_CID_GAIN;
  break;
case CV_CAP_PROP_EXPOSURE:
  sprintf(name, "Exposure");
  capture->control.id = V4L2_CID_EXPOSURE;
  break;
case CV_CAP_PROP_FOCUS:
  sprintf(name, "Focus");
  capture->control.id = V4L2_CID_FOCUS_ABSOLUTE;
  break;
default:
  sprintf(name, "<unknown property string>");
  capture->control.id = property_id;

....

As you can see your error is the default case. But you can modify this file. But i think it is easy to implement, try:

    case CV_CAP_PROP_FPS:
  sprintf(name, "Focus");
  return capture->GetFrameRate() ;

I´ve did the same with CV_CAP_PROP_FOCUS an it works fine.

Best

Surf is looking for eye-catching points. This occurs inter alia by the contrast between the umligenden pixels. I think your buildings have very few of these points.

Here are some Tips:

• Use Grayscale Images (ore cast them to Grayscale)
• Sharp them (Sharp Image = Original Image - Gausian(Original Image))

Maybe you can try a Sobel and add it to the sharpened Images. I think that would give you more matching Keypoints at the roof area and on the buildings.

You should also use a dynamic threshold for matching!. Like:

if((matches2[i][0].distance < 0.56*(matches2[i][1].distance)) && ((int) matches2[i].size()<=2 && (int) matches2[i].size()>0))

You could also try a KNN-Matcher instead of a BF-Matcher.

Hi,

Here you find the exact documentation

http://opencv.willowgarage.com/documentation/cpp/basic_structures.html

uchar* data is the pointer to the image data.

Copy the core243d.dll in your running directory. (In VS usually /Documents/Visual Studio 2010/Projects/<your project="" name="">/<your project="" name="">/) I think there will be more errors of the same kind after you´ve copied the core.dll. So you have to copy all the requested DLLs into this directory.

Try cvCreateCameraCapture(0). Maybe you should use Mat instead of IplImage.

How much Keypoints did you get?

Try Unsharp masking.

-> Sharp Image = Original Image - Gausian(Original Image).

Hi, can you give more details. Like which detector you use. My experience is that the detector requires the most time. Using Windows FREAK is faster then SIFT or Surf. Maybe there´s a issue in the implementation of the Android version. Otherwise the reason could be tha FREAK uses floating point descriptors, like SIFT and SURF.

Hi there, i´m using Ubuntu 12.10 with all dependencies and libv4l2.I want to change the focus of a V4l2 Cam. I´m trieing it with the following code:

.
.
VideoCapture cam;
.
.
cam.set(CV_CAP_PROP_FOCUS ,d);

Capture frames works.

If I use Windows 7 the code works fine.

Same thing with CV_CAP_PROP_FOCUS and many others....

Try this version.

` #include <opencv.hpp> #include <iostream>

using namespace std; using namespace cv;

int main(int argc, char *argv[]) { Mat img = imread("lena.jpg", 1); namedWindow("Lena", CV_WINDOW_AUTOSIZE); imshow("Lena", img); waitKey();

return 0;

} `

Hi, the difference betwenn out and good matches is only the faktor (059 and 0.675), tah gives me the ratio good/out =x, the threshold of x is 0.2. I´ve tried a little with static matches recognition like if good_matches.size()>10, but i figuered out that the ratio way gives me better solution.

Also had a problem using the FLANN Matcher. Just changed

FlannBasedMatcher matcher

to

BFMatcher matcher(NORM_L2,false);