Image retrieval: one-dimensional histogram + Euclidean distance +FLANN+KNN

A text file with a file name of "folder" is generated in the F-drive.

First step: Batch extracts the one-dimensional color histogram of the image and saves it to the featurehists in the. xml

First parameter: The path to the image

Second parameter: the saved. xml

#include <iostream> #include <fstream> #include <string>using namespace std; #include <opencv2\ imgproc\imgproc.hpp> #include <opencv2\core\core.hpp> #include <opencv2\highgui\highgui.hpp>using Namespace cv;//calculates one-dimensional histogram feature Mat hist1d (const mat& SRC); int main (int argc,char* argv[]) {    if (argc!=3) & nbsp;   {        cerr << "wrong Argument!" <<endl;   & nbsp;    return-1;    }    //definition file stream, only can read     ifstream Inputfile (argv[1],ios::in);    if (inputfile)     {        cerr << "File Open erro!" <<endl;        return-1;    }     //reads each row in the file stream and assigns a value to filename. Reads each image and displays     string fileName;    mat image;    mat featurehist; & nbsp  mat FEATUREHISTS;&NBSP;&NBsp  while (Getline (inputfile,filename))     {                image = Imread (filename,1);        //compute one-dimensional histogram features         featurehist = hist1d (image);        //One-dimensional histogram feature for each image by row         featurehists.push_back (featurehist);    }   / /Note Be sure to remember to close the file stream     inputfile.close ();    /* Fifth step. Save image features to an. xml file */    filestorage FS (argv[2],filestorage::write);    fs<< " Featurehists "<<featureHists;    fs.release ();        return 0;} Mat hist1d (const mat& src) {    mat hsv;    //color space conversion bgr2hsv     Cvtcolor (SRC,HSV,CV_BGR2HSV);    //divides h channel into 60 bin    int hbins = 60;     int histsize[] = {Hbins};    //h Value range 0-179    float hranges[]= {0,180};    const float* Ranges [] ={hranges};    mat hist1d,histrow,histrowdst;    //We are based on the first channel of the image. Calculates a one-dimensional histogram and outputs a hist1d of 32f    int channels [] ={0};    calchist (&hsv,1,channels,mat ( ), Hist1d,1,histsize,ranges,true,false);    //to store the histogram feature on one line     histrow= Hist1d.reshape (;  )  //histogram normalization     normalize (HISTROW,HISTROWDST,1,0,NORM_L1);     return HISTROWDST;}

After compiling, go to the command line

Then, a features. xml file appears on the F-disk. The one-dimensional histogram feature of the above image is stored in the inside.

—————————————————————————————————————————————————————————————————————————————
Step two: Extracting one-dimensional color histogram of color cards

 #include < Iostream> #include <string>using namespace std; #include <opencv2\core\core.hpp> #include <opencv2\ Highgui\highgui.hpp> #include <opencv2\imgproc\imgproc.hpp>using namespace cv;int main (int argc,char* argv[] {Mat src = imread (argv[1],1), if (! Src.data) {cout << "No Image" << endl;return-1;} Mat hsv;//Color space Conversion bgr2hsvcvtcolor (SRC,HSV,CV_BGR2HSV);//divide H channel into 60 binint hbins = 60;int histsize[] = {hbins};//h value range 0-179float hranges[]= {0,180};const float* ranges [] ={hranges}; Mat hist1d,histrow,histrowdst;//We calculate a one-dimensional histogram based on the first channel of the image, and the output HIST1D is 32Fint channels [] ={0};calchist (&hsv,1, Channels,mat (), hist1d,1,histsize,ranges,true,false);//The histogram feature is stored in one row for Histrow=hist1d.reshape (n/a); The histogram normalization normalize (HISTROW,HISTROWDST,1,0,NORM_L1); Filestorage FS (argv[2],filestorage::write);//Save HISTROWDST to. xml file FS << argv[3] << histrowdst; Fs.release (); return 0;} 

Individually extracting their one-dimensional color histogram

Then, there were four. xml files in the F-disk, each storing their color histogram information

_______________________________________________________________________________________________________________ _____________________

The third step: retrieving the color histogram of the image using color histogram of colour card

Image retrieval: one-dimensional histogram + Euclidean distance +FLANN+KNN