KORAL_ROS is a ROS based computer vision pipeline, which combines GPU based feature detection, description and matching.
Main features:
- KFAST: FAST(!!) corner detection using AVX2 instructions
- CLATCH: CUDA based LATCH description for features, produces 512-bit binary descriptors.
- CUDAK2NN: Extremely fast Hamming distance based brute force matching.
This repository serves as an extension to KORAL, an extremely fast, highly accurate, scale- and rotation-invariant CPU-GPU cooperative detector-descriptor that uses FAST keypoints and LATCH descriptors, combining it with CUDAK2NN, a super-fast GPU implementation of a brute-force matcher for 512-bit binary descriptors, both originally developed by Kareem Omar. All credits for these amazingly fast kernels go to the original author(s).
In this repository, KORAL and CUDAK2NN have been adapted into a real time framework, where images are read in succession on which detection and matching is performed. The sample code in koral_node.cpp subscribes to two ROS topics 'imageL' and 'imageR', simulating left and right camera views, performs feature extraction on both images and brute force matching between the two views. This is aimed at being a starting point for GPU based vision algorithms for autonomous vehicles: considering that feature detection and matching are some of the biggest bottlenecks, fast GPU based implementations can allow for real-time localization.
Dependencies:
- AVX2 capable CPU
- CUDA capable GPU
- ROS
- OpenCV (for retrieving keypoints and matches)
This code is meant only as an example to get started with, as numerous improvements can be made to the current functionality (example: asynchronous detection for multiple images).
Sample benchmark: (i7-6770HQ, GTX 1080)
| Image resolution | Detection (ms per image) | Matching (ms per pair) |
|---|---|---|
| 640x480 | 3 | 1 |
| 1920x1080 | 7 | 4 |
Note: Matching performance depends on the number of total feature points that need to be evaluated.
Parameters:
Feature extraction:
FeatureDetector detector(scaleFactor, scaleLevels, width, height, maxFeatureCount, fastThreshold);
scaleFactor: Coefficient by which one scale pyramid level is divided to obtain the next.scaleLevels: Number of levels in the scale space.fastThreshold: Threshold intensity difference between pixels (for classifying as a corner).
Feature matching:
FeatureMatcher matcher(matchThreshold, maxFeatureCount);
matchThreshold: A match is returned if the best match between a query vector and a training vector is more thanmatchThresholdnumber of bits better than the second-best match.
[1]
KORAL is licensed under the MIT License : https://opensource.org/licenses/mit-license.php
Copyright(c) 2016 Kareem Omar, Christopher Parker
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files(the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and / or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions :
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
Note again that KORAL is a work in progress. Suggestions and improvements are welcomed.
The FAST detector was created by Edward Rosten and Tom Drummond as described in the 2006 paper by Rosten and Drummond: "Machine learning for high-speed corner detection" Edward Rosten and Tom Drummond https://www.edwardrosten.com/work/rosten_2006_machine.pdf
The FAST detector is BSD licensed:
Copyright(c) 2006, 2008, 2009, 2010 Edward Rosten All rights reserved.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met :
*Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
*Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and / or other materials provided with the distribution.
*Neither the name of the University of Cambridge nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.