init.lua

require 'torch'
require 'paths'
local ffi = require 'ffi'

local pcl = require 'pcl.PointTypes'
local utils = require 'pcl.utils'
require 'pcl.PointCloud'
require 'pcl.Indices'
require 'pcl.IndicesVector'
require 'pcl.Correspondences'
require 'pcl.CloudViewer'
require 'pcl.OpenNI2Stream'
require 'pcl.PCA'
require 'pcl.ICP'
require 'pcl.ICPNL'
require 'pcl.IncrementalRegistration'
require 'pcl.affine'
require 'pcl.primitive'
require 'pcl.filter'
require 'pcl.KdTree'
require 'pcl.Octree'
require 'pcl.NormalEstimation'
require 'pcl.IntegralImageNormalEstimation'
require 'pcl.SIFTKeypoint'
require 'pcl.FPFHEstimation'
require 'pcl.CorrespondenceEstimation'
require 'pcl.SampleConsensusPrerejective'
require 'pcl.BoundaryEstimation'
require 'pcl.OrganizedEdge'
require 'pcl.VFHEstimation'
require 'pcl.SACSegmentation'
require 'pcl.EuclideanClusterExtraction'
require 'pcl.PCLVisualizer'
require 'pcl.PCLPointCloud2'

function pcl.rand(width, height, pointType)
  if pcl.isPointType(height) or type(height) == 'string' then
    pointType = height
    height = 1
  else
    height = height or 1
  end
  local c = pcl.PointCloud(pointType or pcl.PointXYZ, width, height)
  if width > 0 and height > 0 then
    c:pointsXYZ():rand(height, width, 3)
  end
  return c
end

function pcl.isPointCloud(obj)
  return obj and torch.isTypeOf(obj, pcl.PointCloud)
end

function pcl.isPoint(obj)
  if obj and type(obj) == 'cdata' then
    local ok,t = pcall(ffi.typeof, obj)
    return ok and pcl.isPointType(t)
  end
  return false
end

function pcl.loadPCD(filename, pointType)
  local c = pcl.PointCloud(pointType or pcl.PointXYZ)
  c:loadPCDFile(filename)
  return c
end

function pcl.loadPLY(filename, pointType)
  local c = pcl.PointCloud(pointType or pcl.PointXYZ)
  c:loadPLYFile(filename)
  return c
end

function pcl.transformCloud(source, destination, transform)
  utils.check_arg('source', pcl.isPointCloud(source), 'point cloud expected')
  return source:transform(transform, destination)
end

function pcl.getMinMax3D(cloud)
  return cloud:getMinMax3D()
end

function pcl.compute3DCentroid(cloud)
  return cloud:compute3DCentroid()
end

function pcl.computeCovarianceMatrix(cloud, centroid)
  return cloud:computeCovarianceMatrix(centroid)
end

function pcl.copyPointCloud(cloud_in, indices, cloud_out)
  return pcl.PointCloud.copy(cloud_in, indices, cloud_out)
end

return pcl