The goal of this project is to make Point-Cloud Library (PCL) data-structures and algorithms available in Java projects via the Java Native Interface (JNI). Currently supported operating systems are Windows and Linux (64bit architecture only).

Point-Cloud Library version 1.11.1 as well as all of its dependencies need to be installed in order to use pcljava; these dependencies do NOT come packed with the artifacts associated with this project.

• An all-in-one PCL installer exists for Windows; once the installation is complete, the following directories must be added to the PATH environment variable:
  • C:\Program Files\PCL 1.11.1\bin
  • C:\Program Files\PCL 1.11.1\3rdParty\VTK\bin
  • C:\Program Files\CMake\bin
  • C:\Program Files\PCL 1.11.1\3rdParty\Boost\lib
  • C:\Program Files\PCL 1.11.1\3rdParty\FLANN\bin
  • C:\Program Files\PCL 1.11.1\3rdParty\Qhull\bin
  • C:\Program Files\OpenNI2\Redist
• On Linux PCL can be built from source using this guide.

To include the pcljava library in your project, your pom.xml must have the following configuration:

<repositories>
  <repository>
    <id>github-pcljava</id>
    <name>Point-Cloud Library JNI Port Packages</name>
    <url>https://maven.pkg.github.com/vmoglan/pcljava</url>
  </repository>
</repositories>

<dependency>
  <groupId>com.github.vmoglan</groupId>
  <artifactId>pcljava</artifactId>
  <version><!-- e.g. 0.0.1-SNAPSHOT --></version>
</dependency> 

In addition to that, the pom.xml file must be configured to unpack the native dependencies — see the <profiles> section in this example.

The native pcljava library must also be loaded in your project as such:

class Main {
	static {	
		System.loadLibrary("pcljava");
	}
}

• Instantiating a class for which memory is allocated in native code (i.e. a class extending NativeObject):

PointCloud3d cloud = new PointCloud3d(); // creating Java instance
cloud.create(); // allocating memory in the native code

// perform operations on cloud

cloud.dispose(); // freeing the memory allocated in the native code

• Estimating the normal vectors of a three-dimensional point-cloud:

PointCloud3dReaderPly reader = new PointCloud3dReaderPly();
PointCloud3d cloud = reader.read("/path/to/cloud.ply");

NormalEstimation normalEstimation = new NormalEstimation(cloud, 0.03f);
NormalCloud normals = normalEstimation.compute();
		
cloud.dispose();
normals.dispose();

• Sphere detection