Skip to content

Latest commit

 

History

History
134 lines (99 loc) · 4.02 KB

README.md

File metadata and controls

134 lines (99 loc) · 4.02 KB

opencv-installer

Simple application to install OpenCV.

This assumes that the OpenCV artifacts are located in the FRC maven repository at https://first.wpi.edu/FRC/roborio/maven/development or in the local maven repository.

(Artifacts for this project are also available in the above repository)

Link to the maven repo

For now, this is just a CLI app with no GUI.

Supported platforms

This currently supports

  • Windows (32- and 64-bit x86)
  • Mac OS X 10.4 or higher
  • Linux (32- and 64-bit x86, and ARM soft- and hard-float)

Windows notes

Artifacts are installed by default in C:\Users\<user>\OpenCV\. You will need to add the install locations to the PATH environment variable.

Linux notes

JNI and native bindings will be installed in /usr/local/lib and the headers will be installed in /usr/local/include. If you don't have write access to these folders, you can run the installer with

sudo java -Duser.home=$HOME -jar ...

Make sure that /usr/local/lib is on LD_LIBRARY_PATH or the JNI bindings won't be loaded by the JVM.

OS X notes

Similar to Linux, but you probably won't need to use sudo.

Command line arguments

Short name Long name Description Argument
h help Prints the help text
v version Sets the OpenCV version The version in the format x.x.x e.g. 3.1.0
j java Flags the Java API for install. This does not install JNI bindings Install location (optional)
i jni Flags the JNI bindings for install Install location (optional)
s headers Flags the C++ headers for install Install location (optional)
n natives Flags the C++ native libraries for install Install location (optional)
a all Installs all OpenCV artifacts
o overwrite Overwrite already installed files
p platform Download artifacts for a specific platform. They will be located in ./install The platform to download artifacts for

Options for platform

windows-x86
windows-x86_64
osx-x86_64
linux-x86
linux-x86_64
linux-arm
linux-armhf

Usage

java -jar opencv-installer --version <version> --platform <platform> --java <location> --jni <location> --headers <location> --natives <location> --overwrite

Using the installer in Gradle build scripts

buildscript {
  repositories {
    maven {
      url 'https://github.com/WPIRoboticsProjects/opencv-maven/raw/mvn-repo'
    }
  }
  dependencies {
    classpath group: 'edu.wpi.first.wpilib.opencv', name: 'opencv-installer', version: '+'
  }
}

import edu.wpi.first.wpilib.opencv.installer.Installer
import edu.wpi.first.wpilib.opencv.installer.PlatformDetector

def openCvPlatform = PlatformDetector.getPlatform()
def openCvVersion = '3.1.0'

dependencies {
  compile group: 'org.opencv', name: 'opencv-java', version: openCvVersion
  ...
}

task installOpenCvJni(type: Copy) {
  Installer.setVersion(openCvVersion)
  Installer.installJni(openCvPlatform.defaultJniLocation())
  // Or manually set the install location
  Installer.installJni('/usr/local/lib')
}

Using the installer as Java library

First, add

repositories {
  maven {
    url 'https://github.com/WPIRoboticsProjects/opencv-maven/raw/mvn-repo'
  }
}
dependencies {
  compile group: 'edu.wpi.first.wpilib.opencv', name: 'opencv-installer', version: '+'
}

to your build.gradle file.

Then to make sure that OpenCV is installed prior to using any OpenCV code:

import edu.wpi.first.wpilib.opencv.installer.Installer;
import edu.wpi.first.wpilib.opencv.installer.PlatformDetector;
import org.opencv.core.Core;

class Main {

  static {
    Installer.setOpenCvVersion(Core.VERSION);
    Installer.installJni(PlatformDetector.getPlatform().defaultJniLocation());
    System.loadLibrary(Core.NATIVE_LIBRARY_NAME);
  }

}

This will install OpenCV on the current system if the JNI bindings are available for it. If there aren't any JNI bindings, an IOException will be thrown by the call to Installer.installJni()