OpenCV banner matching during sports events

Question

I want to match commercial banners on images during sports events on the walls of arenas. I used this example. If the image is large enough its working quite good (scene_1). But if there is some kind of distortion, or the picture is smaller than some limit (scene_2) then it goes crazy, in addition OpenCV makes matches to points completely unrelated to each other.

Example of a good and failed match:

enter image description here

Is there any way to match only points close together, that make a rectangular shape? Or is there a better method to do matching like this? My biggest problem now is how to make it more precise, because it can get a good match 3 times from 10 occurences (The purpose of this project will be processing whole videos to get some kind of statistics).

This is my code:

#include <stdio.h>
#include <iostream>
#include "opencv2/core/core.hpp"
#include "opencv2/features2d/features2d.hpp"
#include "opencv2/highgui/highgui.hpp"
#include "opencv2/calib3d/calib3d.hpp"
#include "opencv2/nonfree/nonfree.hpp"

using namespace cv;

void readme();

int main( int argc, char** argv )
{
  if( argc != 3 )
  { readme(); return -1; }

  Mat img_object = imread( argv[1], CV_LOAD_IMAGE_GRAYSCALE );
  Mat img_scene = imread( argv[2], CV_LOAD_IMAGE_GRAYSCALE );

  if( !img_object.data || !img_scene.data )
  { std::cout<< " --(!) Error reading images " << std::endl; return -1; }

  //-- Step 1: Detect the keypoints using SURF Detector
  int minHessian = 400;

  SurfFeatureDetector detector( minHessian );

  std::vector<KeyPoint> keypoints_object, keypoints_scene;

  detector.detect( img_object, keypoints_object );
  detector.detect( img_scene, keypoints_scene );

  //-- Step 2: Calculate descriptors (feature vectors)
  SurfDescriptorExtractor extractor;

  Mat descriptors_object, descriptors_scene;

  extractor.compute( img_object, keypoints_object, descriptors_object );
  extractor.compute( img_scene, keypoints_scene, descriptors_scene );

  //-- Step 3: Matching descriptor vectors using FLANN matcher
  FlannBasedMatcher matcher;
  std::vector< DMatch > matches;
  matcher.match( descriptors_object, descriptors_scene, matches );

  double max_dist = 0; double min_dist = 100;

  //-- Quick calculation of max and min distances between keypoints
  for( int i = 0; i < descriptors_object.rows; i++ )
  { double dist = matches[i].distance;
    if( dist < min_dist ) min_dist = dist;
    if( dist > max_dist ) max_dist = dist;
  }

  printf("-- Max dist : %f \n", max_dist );
  printf("-- Min dist : %f \n", min_dist );

  //-- Draw only "good" matches (i.e. whose distance is less than 3*min_dist )
  std::vector< DMatch > good_matches;

  for( int i = 0; i < descriptors_object.rows; i++ )
  { if( matches[i].distance < 3*min_dist )
     { good_matches.push_back( matches[i]); }
  }

  Mat img_matches;
  drawMatches( img_object, keypoints_object, img_scene, keypoints_scene,
               good_matches, img_matches, Scalar::all(-1), Scalar::all(-1),
               vector<char>(), DrawMatchesFlags::NOT_DRAW_SINGLE_POINTS );

  //-- Localize the object
  std::vector<Point2f> obj;
  std::vector<Point2f> scene;

  for( int i = 0; i < good_matches.size(); i++ )
  {
    //-- Get the keypoints from the good matches
    obj.push_back( keypoints_object[ good_matches[i].queryIdx ].pt );
    scene.push_back( keypoints_scene[ good_matches[i].trainIdx ].pt );
  }

  Mat H = findHomography( obj, scene, CV_RANSAC );

  //-- Get the corners from the image_1 ( the object to be "detected" )
  std::vector<Point2f> obj_corners(4);
  obj_corners[0] = cvPoint(0,0); obj_corners[1] = cvPoint( img_object.cols, 0 );
  obj_corners[2] = cvPoint( img_object.cols, img_object.rows ); obj_corners[3] = cvPoint( 0, img_object.rows );
  std::vector<Point2f> scene_corners(4);

  perspectiveTransform( obj_corners, scene_corners, H);

  //-- Draw lines between the corners (the mapped object in the scene - image_2 )
  line( img_matches, scene_corners[0] + Point2f( img_object.cols, 0), scene_corners[1] + Point2f( img_object.cols, 0), Scalar(0, 255, 0), 4 );
  line( img_matches, scene_corners[1] + Point2f( img_object.cols, 0), scene_corners[2] + Point2f( img_object.cols, 0), Scalar( 0, 255, 0), 4 );
  line( img_matches, scene_corners[2] + Point2f( img_object.cols, 0), scene_corners[3] + Point2f( img_object.cols, 0), Scalar( 0, 255, 0), 4 );
  line( img_matches, scene_corners[3] + Point2f( img_object.cols, 0), scene_corners[0] + Point2f( img_object.cols, 0), Scalar( 0, 255, 0), 4 );

  //-- Show detected matches
  imshow( "Good Matches & Object detection", img_matches );

  waitKey(0);
  return 0;
  }

  void readme()
  { std::cout << " Usage: ./SURF_descriptor <img1 - object> <img2 - scene>" << std::endl; }

Have a look at this http://stackoverflow.com/questions/10168686/algorithm-improvement-for-coca-cola-can-shape-recognition — Balaji R, Sep 17 '14 at 06:25
Come on OP, just give me some sweet upvotes, please. I will even mail my source code to you! — James Harper, Sep 22 '14 at 08:29

score 1 · Accepted Answer · answered Sep 19 '14 at 06:57

Try SIFT instead of SURF. Some claim that SURF is more robust, but according to my personal experiences and the recent trend in the CV community SIFT is still the way to go.

Also, you could use some density clustering method to further filter out some outliers, assuming you are only trying to find one object in the target image. A good example of using DBSCAN to improve the detection quality would be

Le, Viet Phuong, et al. "Improving Logo Spotting and Matching for Document Categorization by a Post-Filter Based on Homography." Document Analysis and Recognition (ICDAR), 2013 12th International Conference on. IEEE, 2013.

Their method is very straightforward and easy to implement, I've implemented it myself and found it quite useful for certain tasks. It is just one or two hundred lines of code.

OpenCV banner matching during sports events

1 Answers1