Simple JNI Wrapper for Chris Evans OpenSURF.
(C. Evans, Research Into Robust Visual Features, MSc University of Bristol, 2008.)
This library is distributed under the GNU GPL.
Supports detection of SURF keypoints and computation of descriptors.
- Get Docker for your platform.
If you just want to use OpenSURF in Java you can get a pre-built image. It includes a working copy of OpenSURF and OpenCV 2.4.11 with Oracle Java 8 and Gradle running on Ubuntu Trusty.
-
Pull it from the repo
docker pull stonerworx/opensurf-java
-
Or use it in your Dockerfile
FROM stonerworx/opensurf-java
-
The OpenSURF library is located in /home/javalibs/opensurf-1.0.jar
-
The OpenCV library is located in /home/javalibs/opencv-2411.jar
If you'd like to contribute you can use the Dockerfile to run tests:
-
Build the Dockerfile
docker build -t opensurf-java .
-
Run the test
docker run opensurf-java gradle test
Requires OpenCV to be installed.
-
Install OpenCV (I recommend using MacPorts)
port install opencv +java
-
Copy opencv-2411.jar in /opt/local/share/OpenCV/java to libs/
-
Copy the libopencv_java2411.dylib file in /opt/local/share/OpenCV/java to your java.library.path
-
Build the native library (Use the makefile.mac - You may need to edit some paths in the Makefile)
cd jni
make -
Copy the libopensurf_java.dylib file in libs/native to your java.library.path
-
Build the project
gradle build
-
Add opensurf-1.0.jar as a dependency in your project
Requires the Android NDK to build the libraries.
-
Download the OpenCV Android SDK here: http://sourceforge.net/projects/opencvlibrary/files/opencv-android/
-
Build the native libraries (You need to edit OPENCV_PATH in Android.mk)
cd jni
ndk-build -
Create a jniLibs Folder in your Android project under src/main
-
Copy the armeabi, armeabi-v7a, mips and x86 folders from libs/native to your jniLibs folder
-
Build the project
gradle build
-
Add opensurf-1.0.jar as a dependency in your project
-
See http://stackoverflow.com/a/27421494 for instructions on how to use OpenCV in Android Studio
import com.multicorn.opensurf.Opensurf;
import org.opencv.core.Mat;
import org.opencv.core.MatOfKeyPoint;
import org.opencv.highgui.Highgui;
MatOfKeyPoint keyPoints;
Mat descriptors;
/*
* detect, then describe
*/
Mat imageMat = Highgui.imread(getClass().getResource("/test.jpg").getPath());
keyPoints = Opensurf.detect(imageMat, 5, 4, 2, 0.0004f);
descriptors = Opensurf.describe(imageMat, keyPoints, true);
/*
* detect & describe
*/
keyPoints = new MatOfKeyPoint();
descriptors = new Mat();
Opensurf.detectDescribe(imageMat, keyPoints, descriptors, 5, 4, 2, 0.0004f, true);