I am doing a project on object detection in OpenCV using haar training. I have to mark, in an image, all the places where the object is present. So, I need a code which opens each image so that I can mark the regions.
Similar to the one here:
#include <opencv/cv.h>
#include <opencv/cvaux.h>
#include <opencv/highgui.h>
// for filelisting
#include <stdio.h>
#include <sys/io.h>
// for fileoutput
#include <string>
#include <fstream>
#include <sstream>
#include <dirent.h>
#include <sys/types.h>
using namespace std;
IplImage* image=0;
IplImage* image2=0;
//int start_roi=0;
int roi_x0=0;
int roi_y0=0;
int roi_x1=0;
int roi_y1=0;
int numOfRec=0;
int startDraw = 0;
char* window_name="<SPACE>add <B>save and load next <ESC>exit";
string IntToString(int num)
{
ostringstream myStream; //creates an ostringstream object
myStream << num << flush;
/*
* outputs the number into the string stream and then flushes
* the buffer (makes sure the output is put into the stream)
*/
return(myStream.str()); //returns the string form of the stringstream object
};
void on_mouse(int event,int x,int y,int flag, void *param)
{
if(event==CV_EVENT_LBUTTONDOWN)
{
if(!startDraw)
{
roi_x0=x;
roi_y0=y;
startDraw = 1;
} else {
roi_x1=x;
roi_y1=y;
startDraw = 0;
}
}
if(event==CV_EVENT_MOUSEMOVE && startDraw)
{
//redraw ROI selection
image2=cvCloneImage(image);
cvRectangle(image2,cvPoint(roi_x0,roi_y0),cvPoint(x,y),CV_RGB(255,0,255),1);
cvShowImage(window_name,image2);
cvReleaseImage(&image2);
}
}
int main(int argc, char** argv)
{
char iKey=0;
string strPrefix;
string strPostfix;
string input_directory;
string output_file;
if(argc != 3) {
fprintf(stderr, "%s output_info.txt raw/data/directory/\n", argv[0]);
return -1;
}
input_directory = argv[2];
output_file = argv[1];
/* Get a file listing of all files with in the input directory */
DIR *dir_p = opendir (input_directory.c_str());
struct dirent *dir_entry_p;
if(dir_p == NULL) {
fprintf(stderr, "Failed to open directory %s\n", input_directory.c_str());
return -1;
}
fprintf(stderr, "Object Marker: Input Directory: %s Output File: %s\n", input_directory.c_str(), output_file.c_str());
// init highgui
cvAddSearchPath(input_directory);
cvNamedWindow(window_name,1);
cvSetMouseCallback(window_name,on_mouse, NULL);
fprintf(stderr, "Opening directory...");
// init output of rectangles to the info file
ofstream output(output_file.c_str());
fprintf(stderr, "done.\n");
while((dir_entry_p = readdir(dir_p)) != NULL)
{
numOfRec=0;
if(strcmp(dir_entry_p->d_name, ""))
fprintf(stderr, "Examining file %s\n", dir_entry_p->d_name);
/* TODO: Assign postfix/prefix info */
strPostfix="";
//strPrefix=input_directory;
strPrefix=dir_entry_p->d_name;
//strPrefix+=bmp_file.name;
fprintf(stderr, "Loading image %s\n", strPrefix.c_str());
if((image=cvLoadImage(strPrefix.c_str(),1)) != 0)
{
// work on current image
do
{
cvShowImage(window_name,image);
// used cvWaitKey returns:
// <B>=66 save added rectangles and show next image
// <ESC>=27 exit program
// <Space>=32 add rectangle to current image
// any other key clears rectangle drawing only
iKey=cvWaitKey(0);
switch(iKey)
{
case 27:
cvReleaseImage(&image);
cvDestroyWindow(window_name);
return 0;
case 32:
numOfRec++;
printf(" %d. rect x=%d\ty=%d\tx2h=%d\ty2=%d\n",numOfRec,roi_x0,roi_y0,roi_x1,roi_y1);
//printf(" %d. rect x=%d\ty=%d\twidth=%d\theight=%d\n",numOfRec,roi_x1,roi_y1,roi_x0-roi_x1,roi_y0-roi_y1);
// currently two draw directions possible:
// from top left to bottom right or vice versa
if(roi_x0<roi_x1 && roi_y0<roi_y1)
{
printf(" %d. rect x=%d\ty=%d\twidth=%d\theight=%d\n",numOfRec,roi_x0,roi_y0,roi_x1-roi_x0,roi_y1-roi_y0);
// append rectangle coord to previous line content
strPostfix+=" "+IntToString(roi_x0)+" "+IntToString(roi_y0)+" "+IntToString(roi_x1-roi_x0)+" "+IntToString(roi_y1-roi_y0);
}
else
//(roi_x0>roi_x1 && roi_y0>roi_y1)
{
printf(" hello line no 154\n");
printf(" %d. rect x=%d\ty=%d\twidth=%d\theight=%d\n",numOfRec,roi_x1,roi_y1,roi_x0-roi_x1,roi_y0-roi_y1);
// append rectangle coord to previous line content
strPostfix+=" "+IntToString(roi_x1)+" "+IntToString(roi_y1)+" "+IntToString(roi_x0-roi_x1)+" "+IntToString (roi_y0-roi_y1);
}
break;
}
}
while(iKey!=66);
{
// save to info file as later used for HaarTraining:
// <rel_path>\bmp_file.name numOfRec x0 y0 width0 height0 x1 y1 width1 height1...
if(numOfRec>0 && iKey==66)
{
//append line
/* TODO: Store output information. */
output << strPrefix << " "<< numOfRec << strPostfix <<"\n";
cvReleaseImage(&image);
}
else
{
fprintf(stderr, "Failed to load image, %s\n", strPrefix.c_str());
}
}
}}
output.close();
cvDestroyWindow(window_name);
closedir(dir_p);
return 0;
}
I ran the above code using Visual Studio 10.
When I run the above code it is opening the command window loading all the images (a new named window is also getting opened but getting but it is not staying for me to select the region) and then closing.
When I ran it using start without debugging option in VS10 it is I could see this:
Examining file img (6).jpeg
Loading image img (6).jpeg
Examining file img (6).JPG
Loading image img (6).JPG
Examining file img (7).jpeg
Loading image img (7).jpeg
Examining file img (7).jpg
Loading image img (7).jpg
Examining file img (8).jpeg
Loading image img (8).jpeg
Examining file img (8).jpg
Loading image img (8).jpg
Examining file img (9).jpeg
Loading image img (9).jpeg
Examining file img (9).jpg
Loading image img (9).jpg
Examining file img 1.jpeg
Loading image img 1.jpeg
the object marker fails and cannot load the image
I think that it is skipping the:
if((image=cvLoadImage(strPrefix.c_str(),1)) != 0){.....}
part of the code
There are few currently available object marking tools for Haar training. I use the tool in this link. Its perfectly worked for me. If you want to build your own tool please follow this tutorial. Any way I think when it comes to Haar training, it is waste of time to develop tools to capture coordinates because there are currently available tools for those and it is not the main goal of Haar training. Because Haar training takes more time to train and create cascade XML file. So its better to focus on training.
Related
I want to develop a PhotoBooth and for this, I want to display the last picture I have made with OpenCV in C++ on a Raspberry Pi. So when a new photo is in the source folder, this photo must be loaded. How can I load the newest photo?
My code until now. I load a specific image named "1.bmp", but I want to load with a non-specific name:
int main()
{
Mat image;
image = imread("1.bmp");
namedWindow( "Display", WINDOW_AUTOSIZE );
imshow( "Display", image );
waitKey(0);
return 0;
}
Thanks for your answers.
I am no C++ programmer, but this should give you a pretty good idea of how to do it:
#include <iostream>
#include <string>
#include <sys/types.h>
#include <sys/stat.h>
#include <dirent.h>
int main()
{
struct dirent *drnt;
struct stat stbuf;
DIR *dr;
// Open the current directory
dr=opendir(".");
long long newest=0;
std::string name;
// Iterate through directory entries
while (drnt=readdir(dr)){
// stat the entry to get its type and age
stat(drnt->d_name,&stbuf);
// Check files only - not directories. You may want to do a strcmp() against "BMP" or "bmp" here
if(S_ISREG(stbuf.st_mode)){
long long ns=stbuf.st_mtimespec.tv_sec * 1000000000 + stbuf.st_mtimespec.tv_nsec;
std::cout << drnt->d_name << ": " << ns << std::endl;
// Note this puppy if newer than current newest
if(ns>newest){
newest=ns;
name=drnt->d_name;
}
}
}
closedir(dr);
// Output name of newest
std::cout << name << std::endl;
}
I am need of passing a folder of images as input to Magick++ api. It can be done using mogrify in commandline as shown in post "ImageMagick script to resize folder of images". Reading a single image could be done through api call as
Image image(inputimage)
But how could we do the same for a folder of images? Can anyone help me with the respective api call?
That feature is not included in the Magick++ API. You will need to iterate the directory yourself and then use the Magick++ API to read and write the image. You can find an example on how to iterate through a folder in C/C++ in the following Stack Overflow post: How can I get the list of files in a directory using C or C++?.
I believe you would be responsible of reading the directory. The C library dirent.h is the first thing I think of, but I'm sure there's better C++/system/framework techniques.
#include <iostream>
#include <vector>
#include <dirent.h>
#include <Magick++.h>
int main(int argc, const char * argv[]) {
std::vector<Magick::Image> stack; // Hold images found
DIR * dir_handler = opendir("/tmp/images"); // Open dir
struct dirent * dir_entry;
if (dir_handler != NULL)
{
// Iterate over entries in directory
while ( (dir_entry = readdir(dir_handler)) != NULL ) {
// Only act on regular files
if (dir_entry->d_type == DT_REG) {
// Concatenate path (could be better)
std::string filename("/tmp/images/");
filename += dir_entry->d_name;
// Read image at path
stack.push_back(Magick::Image(filename));
}
}
closedir(dir_handler); // House keeping
} else {
// Handle DIR error
}
// Append all images into single montage
Magick::Image output;
Magick::appendImages(&output, stack.begin(), stack.end());
output.write("/tmp/all.png");
return 0;
}
There's also ExpandFilenames(int *,char ***) in the MagickCore library.
// Patterns to scan
int pattern_count = 1;
// First pattern
char pattern[PATH_MAX] = "/tmp/images/*.png";
// Allocate memory for list of patterns
char ** dir_pattern = (char **)MagickCore::AcquireMagickMemory(sizeof(char *));
// Assign first pattern
dir_pattern[0] = pattern;
// Expand patterns
Magick::MagickBooleanType ok;
ok = MagickCore::ExpandFilenames(&pattern_count, &dir_pattern);
if (ok == Magick::MagickTrue) {
std::vector<Magick::Image> stack;
// `pattern_count' now holds results count
for ( int i = 0; i < pattern_count; ++i) {
// `dir_pattern' has been re-allocated with found results
std::string filename(dir_pattern[i]);
stack.push_back(Magick::Image(filename));
}
Magick::Image output;
Magick::appendImages(&output, stack.begin(), stack.end());
output.write("/tmp/all.png");
} else {
// Error handle
}
So I have played around in OpenCV a bunch before and never run into this problem. I am implementing a MeanShift algorithm and trying to do it on video devices, images, and videos. Devices and images work; however, no matter what I try, when I run VideoCapture on my filename (whether setting it in the Constructor or using the VideoCapture::open() method, and whether local or with a full path) I always get stuck in my error check.
Thoughts? Ideas? code below. running in Visual Studio 2012
#include "opencv2\highgui\highgui.hpp"
#include "opencv2\core\core.hpp"
#include "opencv2\opencv.hpp"
#include "opencv2\video\video.hpp"
#include <string>
using cv::Mat;
using std::string;
enum Filetype{Image, Video};
int main(int argc, char* argv[])
{
string filename = "short_front.avi";// "C:\\Users\\Jonathan\\Videos\\short_front.mp4"; //"hallways.jpg";
Mat cv_image; //convert to unsigned char * with data
Mat filtImage_;
Mat segmImage_;
Mat whiteImage_;
cv::VideoCapture vid;
vid.open("C:/Users/Jonathan/Desktop/TestMeanShift/TestMeanShift/short_front.avi");
cv::waitKey(1000);
if ( !vid.isOpened() ){
throw "Error when reading vid";
cv::waitKey(0);
return -1;
}
// cv_image = cv::imread(filename);//, CV_LOAD_IMAGE_COLOR);
// if(! cv_image.data){
// std::cerr << "Image Failure: " << std::endl;
// system("pause");
// return -1;
// }
//Mat cv_image_gray;
//cv::cvtColor(cv_image,cv_image_gray,CV_RGB2GRAY);
for (;;)
{
vid >> cv_image;
if ( !cv_image.data)
continue;
cv::imshow("Input",cv_image); //add a normal window here to resizable
}
EDIT: This is a distinct problem from the one listed here because it deals with a specific corner case: VideoCapture and ImageCapture both work, only not VideoCapture with a file. When it doesn't work, the code runs properly, except that the "video" it creates is incomplete as it didn't open properly. Therefore, as the code above does not crash in compile time or run time, the only indicator is bad output (6KB video output file). If you are having issues not with the corner case I am describing but general issues with the above functions in OpenCV, the aforementioned link could help you.
I am trying to run the ezsift library example. The example has name "image_match.cpp".
Here is the code
image_match.cpp
#include <iostream>
#include <vector>
#include <stdio.h>
#include <stdlib.h>
#include <ctype.h>
#include "../ezsift.h"
using namespace std;
#define USE_FIX_FILENAME 0
int main(int argc, char ** argv)
{
#if USE_FIX_FILENAME
char * file1 = "img1.pgm";
char * file2 = "img2.pgm";
#else
if (argc != 3)
{
printf("Please input two image filenames.\n");
printf("usage: image_match img1 img2\n");
return -1;
}
char file1[255];
char file2[255];
memcpy(file1, argv[1], sizeof(char) * strlen(argv[1]));
file1[strlen(argv[1])] = 0;
memcpy(file2, argv[2], sizeof(char) * strlen(argv[2]));
file2[strlen(argv[2])] = 0;
#endif
// Read two input images
ImageObj<uchar> image1, image2;
if(image1.read_pgm(file1) != 0)
{
printf("Failed to open input image1!\n");
return -1;
}
if(image2.read_pgm(file2) != 0)
{
printf("Failed to open input image2!\n");
return -1;
}
printf("Image 1 loaded. Image size: %d x %d\n", image1.w, image1.h);
printf("Image 2 loaded. Image size: %d x %d\n", image2.w, image2.h);
// Double the original image as the first octive.
double_original_image(true);
// Detect keypoints
list<SiftKeypoint> kpt_list1, kpt_list2;
printf("\nSIFT detection on image 1 ...\n");
sift_cpu(image1, kpt_list1, true);
printf("# keypoints in image1: %d\n", kpt_list1.size());
printf("\nSIFT detection on image 2 ...\n");
sift_cpu(image2, kpt_list2, true);
printf("# keypoints in image2: %d\n", kpt_list2.size());
// Save keypoint list, and draw keypoints on images.
char filename[255];
sprintf(filename, "s_A_keypoints.ppm");
draw_keypoints_to_ppm_file(filename, image1, kpt_list1);
export_kpt_list_to_file("s_A_keypoints.key", kpt_list1, true);
sprintf(filename, "s_B_keypoints.ppm");
draw_keypoints_to_ppm_file(filename, image2, kpt_list2);
export_kpt_list_to_file("s_B_keypoints.key", kpt_list2, true);
// Match keypoints.
list<MatchPair> match_list;
match_keypoints(kpt_list1, kpt_list2, match_list);
// Draw result image.
sprintf(filename, "s_A_B_matching.ppm");
draw_match_lines_to_ppm_file(filename, image1, image2, match_list);
printf("# of matched keypoints: %d\n", match_list.size());
return 0;
}
ezsift.h
#ifndef EZSIFT_H
#define EZSIFT_H
#include "util/image.h"
#include "util/img_io.h"
#include <vector>
#include <list>
.
.
.
#endif
#define DEGREE_OF_DESCRIPTORS (128)
// Enable doubling of original image.
void double_original_image(bool doubleFirstOctave);
// Detect keypoints and extract descriptor.
int sift_cpu(
const ImageObj<uchar> &image,
std::list<SiftKeypoint> & kpt_list,
bool bExtractDescriptors);
// Match keypoints from two keypoint lists.
int match_keypoints(
std::list<SiftKeypoint> & kpt_list1,
std::list<SiftKeypoint> & kpt_list2,
std::list<MatchPair> & match_list);
.
.
.
#endif
(I will not include all the code here, I just wanted to give an idea. It is ease to download the source and see the files that I am talking about)
Also there is a "ezsift.cpp" file with more files in the "util" folder
I'm trying to run it from the terminal. I go to the directory and type "gcc image_match.cpp" but is says that double_original_image, sift_cpu (and all other functions in the header file) are not defined.
more specifically is gives this error:
image_match.cpp:(.text+0x1c7): undefined reference to `double_original_image(bool)'
image_match.cpp:(.text+0x20d): undefined reference to `sift_cpu(ImageObj<unsigned char> const&, std::list<_SiftKeypoint, std::allocator<_SiftKeypoint> >&, bool)'
(these are not the only error of course. It gives same error for all the functions)
I then tried "gcc ../ezsift.cpp image_match.cpp" but the same error.
What am I doing wrong?? I am running ubuntu 14.04.
(expending Cameron's comment into a tested answer)
You need to compile both ezsift.cpp and img_io.cpp with your program (image_match.cpp), while providing the location of the header files (-I\path\to\directory). Here is a working snippet, assuming you are in the examples directory:
g++ image_match.cpp ../ezsift.cpp ../util/img_io.cpp -I.. -o image_match
Deal all,
I need to decode an animated gif format picture into some bitmap files in MFC2010. Is there any library to decode a gif picture? I cannot use GDIPlus because the program has to run on windows XP. I do appreciate if someone provides me with a library, Activex, dll or anything similar.
Many Thanks,
Shervin Zargham
It's pretty simple using ImageMagick's C++ API (Magick++) :
/* list of Image to store the GIF's frames */
std::vector<Magick::Image> imageList;
/* read all the frames of the animated GIF */
Magick::readImages( &imageList, "animated.gif" );
/* optionnally coalesce the frame sequence depending on the expected result */
Magick::coalesceImages( &imageList, imageList.begin(), imageList.end());
/* store each frame in a separate BMP file */
for(unsigned int i = 0; i < imageList.size(); ++i) {
std::stringstream ss;
ss << "frame" << i << ".bmp";
imageList[i].write(ss.str());
}
WIC (included in Vista, available for XP) offers CLSID_WICGifDecoder, a COM component.
Try this using ImageMagick's C++ API (Magick++) ,tested on VS210:
#include <Magick++.h>
#include <string>
#include <iostream>
#include <list>
using namespace std;
using namespace Magick;
void kk(char * nombre, char *ext)
{
/* list of Image to store the GIF's frames */
std::list<Magick::Image> imageList;
/* read all the frames of the animated GIF */
Magick::readImages( &imageList, nombre );
/* compone las diferencias para obtener los cuadros reales */
Magick::coalesceImages(&imageList,imageList.begin( ),imageList.end( ));
/* store each frame in a separate BMP file */
list <Magick::Image>::iterator it;
int i=1;
for ( it = imageList.begin( ); it != imageList.end( ); it++ , i++)
{
std::string name = "frame" + to_string((_Longlong)(i)) + ext ;
it->write(name);
}
}
int main( int /*argc*/, char ** argv)
{
// Initialize ImageMagick install location for Windows
InitializeMagick(*argv);
try {
kk("luni0.gif", ".png"); // using ".bmp", ".jpg", ".png", OK
return 0;
}
catch( exception &error_ )
{
cout << "Caught exception: " << error_.what() << endl;
return 1;
}
}
It's been a long time, but I recall once using OleLoadPicture to open GIF and PNG files on old versions of Windows, though the documentation seems to suggest that it's only for BMP, ICO, and WMF.