oym1994 / imagefeaturesandmatching

Image features and related matching methods

Image Features Matching

motivation

Design an abstract class where multiple image features, descriptors and matching methods can be chosen from the config file, including traditional and deep learning based methods, for some visual feature based tasks, such as visual slam, AR/VR and object tracking.

Supportted image features: BRISK, ORB, AKAZE, SIFT, SURF, CONTOUR, SuperPoint (with TensorRT accelerated)

Supportted descriptor types: BRIEF, ORB, FREAK, AKAZE, SIFT, BRISK, SURF, SuperPoint (with TensorRT accelerated)

Supported matching methods: BF_HAMMING, BF_L2, SuperGlue (with TensorRT accelerated)

Dependencies for Running Locally

• cmake >= 2.8
  • All OSes: click here for installation instructions
• make >= 4.1 (Linux, Mac), 3.81 (Windows)
  • Linux: make is installed by default on most Linux distros
  • Mac: install Xcode command line tools to get make
  • Windows: Click here for installation instructions
• OpenCV >= 4.1
  • This must be compiled from source using the -D OPENCV_ENABLE_NONFREE=ON cmake flag for testing the SIFT and SURF detectors.
  • The OpenCV 4.1.0 source code can be found here
• gcc/g++ >= 5.4
  • Linux: gcc / g++ is installed by default on most Linux distros
  • Mac: same deal as make - install Xcode command line tools
  • Windows: recommend using MinGW
• optional for USE_DEEP_FEATURES
  • TensorRT 8.2.3
  • Libtorch 1.8

Basic Build Instructions

1. Clone this repo.
2. Make a build directory in the top level directory: mkdir build && cd build
3. Compile: cmake .. && make
4. Run it: ./image_matching [keypoints type] [descriptor type].

---

Reference

SFND_2D_Feature_Tracking

feature detection

Image Matching from Handcrafted to Deep Features: A Survey

Image Matching Across Wide Baselines: From Paper to Practice

Image Registration Techniques: A Survey

ORB

SIFT

SuperPoint

SuperGlue

BRISK

BRIEF

FREAK

AKAZE

imagefeaturesandmatching's Issues

would you provide superglue.pth to onnx scripts

when I using python tensorrt api to convert onnx to trt model, the points0 is dynamic,but points1 is not.I want to align to your onnx model

SuperGlue can't be used

hello oym1994
There's some question about the SuperGlue!
First, the ".engine" of SuperGlue can't be generated;
Second, the matcher of SuplerGlue isn‘’t defined；
so, is there any code to be added, thank you!