doc/tutorials/features2d/feature_homography/feature_homography.markdown

   1 Features2D + Homography to find a known object {#tutorial_feature_homography}
   2 ==============================================
   3
   4 Goal
   5 ----
   6
   7 In this tutorial you will learn how to:
   8
   9 -   Use the function @ref cv::findHomography to find the transform between matched keypoints.
  10 -   Use the function @ref cv::perspectiveTransform to map the points.
  11
  12 \warning You need the <a href="https://github.com/opencv/opencv_contrib">OpenCV contrib modules</a> to be able to use the SURF features
  13 (alternatives are ORB, KAZE, ... features).
  14
  15 Theory
  16 ------
  17
  18 Code
  19 ----
  20
  21 @add_toggle_cpp
  22 This tutorial code's is shown lines below. You can also download it from
  23 [here](https://github.com/opencv/opencv/tree/3.4/samples/cpp/tutorial_code/features2D/feature_homography/SURF_FLANN_matching_homography_Demo.cpp)
  24 @include samples/cpp/tutorial_code/features2D/feature_homography/SURF_FLANN_matching_homography_Demo.cpp
  25 @end_toggle
  26
  27 @add_toggle_java
  28 This tutorial code's is shown lines below. You can also download it from
  29 [here](https://github.com/opencv/opencv/tree/3.4/samples/java/tutorial_code/features2D/feature_homography/SURFFLANNMatchingHomographyDemo.java)
  30 @include samples/java/tutorial_code/features2D/feature_homography/SURFFLANNMatchingHomographyDemo.java
  31 @end_toggle
  32
  33 @add_toggle_python
  34 This tutorial code's is shown lines below. You can also download it from
  35 [here](https://github.com/opencv/opencv/tree/3.4/samples/python/tutorial_code/features2D/feature_homography/SURF_FLANN_matching_homography_Demo.py)
  36 @include samples/python/tutorial_code/features2D/feature_homography/SURF_FLANN_matching_homography_Demo.py
  37 @end_toggle
  38
  39 Explanation
  40 -----------
  41
  42 Result
  43 ------
  44
  45 -   And here is the result for the detected object (highlighted in green). Note that since the homography is estimated with a RANSAC approach,
  46     detected false matches will not impact the homography calculation.
  47
  48     ![](images/Feature_Homography_Result.jpg)