VS2010+Opencv2.3.1，例程實現

020 Template Matching

成功程式：

/**

 * @file MatchTemplate_Demo.cpp

 * @brief Sample code to use the function MatchTemplate

 * @author OpenCV team

 */

#include

#include"opencv2/imgproc/imgproc.hpp"

#include<iostream>

#include<stdio.h>

usingnamespace std;

usingnamespace cv;

/// Global Variables

Mat img; Mat templ; Mat result;

char* image_window = "Source Image"

char* result_window = "Result window";

int match_method;

int max_Trackbar = 5;

/// Function Headers

void MatchingMethod( int, void* );

/**

 * @function main

 */

int main( int argc, char** argv )

{

  /// Load image and template

  img = imread( "H:\\Image\\image0.jpg" , 1 ); //

  templ = imread( "H:\\Image\\image0010.jpg" , 1 );

  /// Create windows

  namedWindow( image_window, CV_WINDOW_AUTOSIZE );

  namedWindow( result_window, CV_WINDOW_AUTOSIZE );

  /// Create Trackbar

  char* trackbar_label = "Method: \n 0: SQDIFF \n 1: SQDIFF NORMED \n 2: TM CCORR \n 3: TM CCORR NORMED \n 4: TM COEFF \n 5: TM COEFF NORMED";

  createTrackbar( trackbar_label, image_window, &match_method, max_Trackbar, MatchingMethod );

  MatchingMethod( 0, 0 );

  waitKey(0);

  return 0;

}

/**

 * @function MatchingMethod

 * @brief Trackbar callback

 */

void MatchingMethod( int, void* )

{

  /// Source image to display

  Mat img_display;

  img.copyTo( img_display );

  /// Create the result matrix

  int result_cols =  img.cols - templ.cols + 1;

  int result_rows = img.rows - templ.rows + 1;   

  result.create( result_cols, result_rows, CV_32FC1 );

  /// Do the Matching and Normalize

  matchTemplate( img, templ, result, match_method );

  normalize( result, result, 0, 1, NORM_MINMAX, -1, Mat() );

  /// Localizing the best match with minMaxLoc

  double minVal; double maxVal; Point minLoc; Point maxLoc;

  Point matchLoc;

  minMaxLoc( result, &minVal, &maxVal, &minLoc, &maxLoc, Mat() );

  /// For SQDIFF and SQDIFF_NORMED, the best matches are lower values. For all the other methods, the higher the better

  if( match_method  == CV_TM_SQDIFF || match_method == CV_TM_SQDIFF_NORMED )

    { matchLoc = minLoc; }

  else  

    { matchLoc = maxLoc; }

  /// Show me what you got

  rectangle( img_display, matchLoc, Point( matchLoc.x + templ.cols , matchLoc.y + templ.rows ), Scalar::all(0), 2, 8, 0 ); 

  rectangle( result, matchLoc, Point( matchLoc.x + templ.cols , matchLoc.y + templ.rows ), Scalar::all(0), 2, 8, 0 ); 

  imshow( image_window, img_display );

  imshow( result_window, result );

  return;

}

原圖

 

 模板圖：

執行結果：

看打框框的就表示程式找到了模板圖的目標了,//左上角的圖表示的是相似度圖，最亮的就表示相似度最高！

程式詳解：

Explanation 

    1. Declare some global variables, such as the image, template and result matrices, as well as the match method and 

       the window names: 

       Mat  img;  Mat templ;   Mat  result; 

       char*  image_window    = "Source   Image"; 

       char*  result  window   =  "Result  window"; 

       int  match_method; 

       int  max_Trackbar   =  5 ; 

   2.  Load the source image and template: 

       img  = imread(   argv[1],   1 ); 

       templ  = imread(   argv[2],   1  ); 

   3.  Create the windows to show the results: 

       namedWindow(    image_window,   CV_WINDOW_AUTOSIZE     ); 

       namedWindow(    result_window,    CV_WINDOW_AUTOSIZE     ); 

   4.  Create the Trackbar to enter the kind of matching method to be used.  When a change is detected the callback 

       function MatchingMethod is called. 

       char*  trackbar_label    =  "Method:   \n 0:  SQDIFF  \n  1:  SQDIFF  NORMED   \n 2:  TM  CCORR  \n  3: TM  CCORR  NORMED   \n 

       createTrackbar(    trackbar_label,     image_window,   &match_method,     max_Trackbar,    MatchingMethod    ); 

   5.  Wait until user exits the program. 

       waitKey(0); 

       return  0 ; 

   6.  Let’s check out the callback function. First, it makes a copy of the source image: 

       Mat  img_display; 

       img.copyTo(   img_display    ); 

   7.  Next, it creates the result matrix that will store the matching results for each template location. Observe in detail 

       the size of the result matrix (which matches all possible locations for it) 

       int  result  cols  =   img.cols   -  templ.cols  +  1; 

       int  result  rows  =  img.rows   -  templ.rows  +  1; 

   8.  Perform the template matching operation: 

       matchTemplate(    img,  templ,  result,   match_method    ); 

       the arguments are naturally the input image I, the template T, the result R and the match_method (given by the 

       Trackbar) 

   9.  We normalize the results: 

       normalize(   result,   result,  0 , 1,  NORM  MINMAX,   -1,  Mat()  ); 

   10. We localize the minimum and maximum values in the result matrix R by using minMaxLoc. //怪不得，最亮的就表示相似度最高！

       double  minVal;   double  maxVal;   Point  minLoc;   Point  maxLoc; 

       Point  matchLoc;

       minMaxLoc(    result,  &minVal,   &maxVal,  &minLoc,   &maxLoc,   Mat()  ); 

      the function calls as arguments: 

          •  result: The source array 

          •  &minVal and &maxVal: Variables to save the minimum and maximum values in result 

          •  &minLoc and &maxLoc: The Point locations of the minimum and maximum values in the array. 

          •  Mat(): Optional mask 

  11. For the first two methods ( CV_SQDIFF and CV_SQDIFF_NORMED ) the best match are the lowest values. 

      For all the others, higher values represent better matches. So, we save the corresponding value in the matchLoc 

      variable: 

         { matchLoc   = minLoc;   } 

      else 

         { matchLoc   = maxLoc;   } 

  12. Display the source image and the result matrix. Draw a rectangle around the highest possible matching area: 

       rectangle(   result,  matchLoc,   Point(   matchLoc.x  +  templ.cols   , matchLoc.y   +  templ.rows   ), Scalar ::all(0),   2,  8 ,  0 

       imshow(  image_window,    img_display   ); 

       imshow(  result_window,    result  );