BodyScanner is an easy to use, open source Body Scanning solution running on top of OpenNI (http://www.openni.org) and the Point Cloud Library (http://www.pointclouds.org).
BodyScanner is in it's very first moments of life! It is not yet ready for general consumption.
To build, type the following in this directory:
mkdir build cd build cmake .. make
This will build all the subprojects.
Expand the mark1_pcd.tar.bz2 archive into the BodyScanner/recordings/mark1_pcd folder. This folder is ignored by git, so it will not be committed.
When our programs are run, they will be called from their directories inside BodyScanner/build/subprojectXYZ/ So when loading a PCD file, go up two directories, like this: "../../recordings/mark1_pcd/1316652689.744386960.pcd"
To build a triangle mesh from a scan, type the following in the build/surface-reconstructor directory
make ./surface-reconstructor -s ../../recordings/mark1_pcd/1316652689.744386960.pcd # read and smooth point cloud ./surface-reconstructor -r # reconstruct the surface ./vtk2obj.py scan.vtk scan.obj # convert into a format that meshlab recognizes (and re-orient faces)
To view the scan.obj that was created, install meshlab:
sudo apt-get install meshlab
And run it:
meshlab scan.obj &
Grab the SDK and install it with these commands:
wget http://luffel.org/2011/body-scanner/fbx20122_fbxsdk_linux.tar.gz tar -xzvf fbx20122_fbxsdk_linux.tar.gz chmod u+x fbx20122_fbxfilesdk_linux sudo ./fbx20122_fbxfilesdk_linux /usr
Here's how to build a fully rigged mesh from the skeleton and point cloud data. Not that this totally works.
DATA=recordings/manohar OUTPUT=records/manohar_files PCD=$DATA/manohar_00001.pcd YAML=$DATA/manohar_skeleton_00001.yaml WEI=$OUTPUT/00001.w OBJ=$OUTPUT/00001.obj FBX=$OUTPUT/00001.fbx mkdir $OUTPUT
segmentation/build/sharpSegmentation $PCD $YAML $WEI surface-reconstructor/build/surface-reconstructor $PCD $OBJ python MakeRig.py $YAML $OBJ $WEI $FBX