Conversion in Point Cloud Library from proprietary format to float; no matching function found

I have what I think is a fairly basic question about a conversion in the point cloud library from a proprietry data type to float. My issue I think comes from a lack of experience with templates, data types and C++ in general.

The function that should perform this conversion is copyToFloatArray, defined in the documentation as:

virtual void pcl::DefaultPointRepresentation< FPFHSignature33 >::copyToFloatArray ( const FPFHSignature33 & p, float * out

) const [inline, virtual]

Copy point data from input point to a float array.

This method must be overriden in all subclasses.

Parameters:

p The input point

out A pointer to a float array.

Implements pcl::PointRepresentation< FPFH开发者_开发问答Signature33 >.

I have attempted to implement it as follows:

pcl::PointCloud<pcl::FPFHSignature33>::Ptr fpfhs (new pcl::PointCloud<pcl::FPFHSignature33> ());
// populate fpfh...
float **myArray;
pcl::DefaultPointRepresentation< pcl::FPFHSignature33 >::copyToFloatArray ( &fpfhs, &**myArray);

The error that's thrown up at compilation time is as follows:

/home/bc/PCL/pcd_read.cpp: In function ‘int main(int, char**)’:
/home/bc/PCL/pcd_read.cpp:68: error: no matching function for call to 
‘pcl::DefaultPointRepresentation<pcl::FPFHSignature33>::copyToFloatArray(boost::shared_ptr<pcl::PointCloud<pcl::FPFHSignature33> >*, float*)’
/usr/include/pcl-1.2/pcl/point_representation.h:254: note: candidates are: 
virtual void pcl::DefaultPointRepresentation<pcl::FPFHSignature33>::copyToFloatArray(const pcl::FPFHSignature33&, float*) const
make[2]: *** [CMakeFiles/pcd_read.dir/pcd_read.cpp.o] Error 1
make[1]: *** [CMakeFiles/pcd_read.dir/all] Error 2
make: *** [all] Error 2

It appears that the issue is in the first argument being passed to the function, but I can't seem to create a const FPFHSignature33 & p object.

Does anyone have any idea if these suspicions are correct, and if so what direction I might take to start to resolve the issue?

Thanks for any help.

First, you need to allocate memory in your floatArray. Then, you need to derefence the fpfhs ptr using *. finally, there is no need for a double pointer array there.

Here is the corrected code :

pcl::PointCloud<pcl::FPFHSignature33>::Ptr fpfhs (new pcl::PointCloud<pcl::FPFHSignature33> ());
// populate fpfh...
float *myArray = new float[ 3* point_count ];
pcl::DefaultPointRepresentation< pcl::FPFHSignature33 >::copyToFloatArray ( *fpfhs, myArray);