doc/tutorials/features2d/feature_homography/feature_homography.markdown

   1 Features2D + Homography to find a known object {#tutorial_feature_homography}
   2 ==============================================
   3
   4 @prev_tutorial{tutorial_feature_flann_matcher}
   5 @next_tutorial{tutorial_detection_of_planar_objects}
   6
   7 Goal
   8 ----
   9
  10 In this tutorial you will learn how to:
  11
  12 -   Use the function @ref cv::findHomography to find the transform between matched keypoints.
  13 -   Use the function @ref cv::perspectiveTransform to map the points.
  14
  15 \warning You need the <a href="https://github.com/opencv/opencv_contrib">OpenCV contrib modules</a> to be able to use the SURF features
  16 (alternatives are ORB, KAZE, ... features).
  17
  18 Theory
  19 ------
  20
  21 Code
  22 ----
  23
  24 @add_toggle_cpp
  25 This tutorial code's is shown lines below. You can also download it from
  26 [here](https://github.com/opencv/opencv/tree/3.4/samples/cpp/tutorial_code/features2D/feature_homography/SURF_FLANN_matching_homography_Demo.cpp)
  27 @include samples/cpp/tutorial_code/features2D/feature_homography/SURF_FLANN_matching_homography_Demo.cpp
  28 @end_toggle
  29
  30 @add_toggle_java
  31 This tutorial code's is shown lines below. You can also download it from
  32 [here](https://github.com/opencv/opencv/tree/3.4/samples/java/tutorial_code/features2D/feature_homography/SURFFLANNMatchingHomographyDemo.java)
  33 @include samples/java/tutorial_code/features2D/feature_homography/SURFFLANNMatchingHomographyDemo.java
  34 @end_toggle
  35
  36 @add_toggle_python
  37 This tutorial code's is shown lines below. You can also download it from
  38 [here](https://github.com/opencv/opencv/tree/3.4/samples/python/tutorial_code/features2D/feature_homography/SURF_FLANN_matching_homography_Demo.py)
  39 @include samples/python/tutorial_code/features2D/feature_homography/SURF_FLANN_matching_homography_Demo.py
  40 @end_toggle
  41
  42 Explanation
  43 -----------
  44
  45 Result
  46 ------
  47
  48 -   And here is the result for the detected object (highlighted in green). Note that since the homography is estimated with a RANSAC approach,
  49     detected false matches will not impact the homography calculation.
  50
  51     ![](images/Feature_Homography_Result.jpg)