feature_estimation.h
Go to the documentation of this file.
00001 #ifndef FEATURE_ESTIMATION_H
00002 #define FEATURE_ESTIMATION_H
00003 
00004 #include "typedefs.h"
00005 
00006 #include <pcl/io/io.h>
00007 #include <pcl/features/normal_3d.h>
00008 #include <pcl/keypoints/sift_keypoint.h>
00009 #include <pcl/features/fpfh.h>
00010 #include <pcl/features/vfh.h>
00011 #include <pcl/search/kdtree.h>
00012 /* Use NormalEstimation to estimate a cloud's surface normals 
00013  * Inputs:
00014  *   input
00015  *     The input point cloud
00016  *   radius
00017  *     The size of the local neighborhood used to estimate the surface
00018  * Return: A pointer to a SurfaceNormals point cloud
00019  */
00020 SurfaceNormalsPtr
00021 estimateSurfaceNormals (const PointCloudPtr & input, float radius)
00022 {
00023   pcl::NormalEstimation<PointT, NormalT> normal_estimation;
00024   normal_estimation.setSearchMethod (pcl::search::Search<PointT>::Ptr (new pcl::search::KdTree<PointT>));
00025   normal_estimation.setRadiusSearch (radius);
00026   normal_estimation.setInputCloud (input);
00027   SurfaceNormalsPtr normals (new SurfaceNormals);
00028   normal_estimation.compute (*normals);
00029 
00030   return (normals);
00031 }
00032 
00033 /* Use SIFTKeypoint to detect a set of keypoints
00034  * Inputs:
00035  *   points
00036  *     The input point cloud
00037  *   normals
00038  *     The input surface normals
00039  *   min_scale
00040  *     The smallest scale in the difference-of-Gaussians (DoG) scale-space
00041  *   nr_octaves
00042  *     The number of times the scale doubles in the DoG scale-space
00043  *   nr_scales_per_octave
00044  *     The number of scales computed for each doubling
00045  *   min_contrast
00046  *     The minimum local contrast that must be present for a keypoint to be detected
00047  * Return: A pointer to a point cloud of keypoints
00048  */
00049 PointCloudPtr
00050 detectKeypoints (const PointCloudPtr & points, const SurfaceNormalsPtr & normals,
00051                  float min_scale, int nr_octaves, int nr_scales_per_octave, float min_contrast)
00052 {
00053   pcl::SIFTKeypoint<PointT, pcl::PointWithScale> sift_detect;
00054   sift_detect.setSearchMethod (pcl::search::Search<PointT>::Ptr (new pcl::search::KdTree<PointT>));
00055   sift_detect.setScales (min_scale, nr_octaves, nr_scales_per_octave);
00056   sift_detect.setMinimumContrast (min_contrast);
00057   sift_detect.setInputCloud (points);
00058   pcl::PointCloud<pcl::PointWithScale> keypoints_temp;
00059   sift_detect.compute (keypoints_temp);
00060   PointCloudPtr keypoints (new PointCloud);
00061   pcl::copyPointCloud (keypoints_temp, *keypoints);
00062 
00063   return (keypoints);
00064 }
00065 
00066 /* Use FPFHEstimation to compute local feature descriptors around each keypoint
00067  * Inputs:
00068  *   points
00069  *     The input point cloud
00070  *   normals
00071  *     The input surface normals
00072  *   keypoints
00073  *     A cloud of keypoints specifying the positions at which the descriptors should be computed
00074  *   feature_radius
00075  *     The size of the neighborhood from which the local descriptors will be computed 
00076  * Return: A pointer to a LocalDescriptors (a cloud of LocalDescriptorT points)
00077  */
00078 LocalDescriptorsPtr
00079 computeLocalDescriptors (const PointCloudPtr & points, const SurfaceNormalsPtr & normals, 
00080                          const PointCloudPtr & keypoints, float feature_radius)
00081 {
00082   pcl::FPFHEstimation<PointT, NormalT, LocalDescriptorT> fpfh_estimation;
00083   fpfh_estimation.setSearchMethod (pcl::search::Search<PointT>::Ptr (new pcl::search::KdTree<PointT>));
00084   fpfh_estimation.setRadiusSearch (feature_radius);
00085   fpfh_estimation.setSearchSurface (points);  
00086   fpfh_estimation.setInputNormals (normals);
00087   fpfh_estimation.setInputCloud (keypoints);
00088   LocalDescriptorsPtr local_descriptors (new LocalDescriptors);
00089   fpfh_estimation.compute (*local_descriptors);
00090 
00091   return (local_descriptors);
00092 }
00093 
00094 /* Use VFHEstimation to compute a single global descriptor for the entire input cloud
00095  * Inputs:
00096  *   points
00097  *     The input point cloud
00098  *   normals
00099  *     The input surface normals
00100  * Return: A pointer to a GlobalDescriptors point cloud (a cloud containing a single GlobalDescriptorT point)
00101  */
00102 GlobalDescriptorsPtr
00103 computeGlobalDescriptor (const PointCloudPtr & points, const SurfaceNormalsPtr & normals)
00104 {
00105   pcl::VFHEstimation<PointT, NormalT, GlobalDescriptorT> vfh_estimation;
00106   vfh_estimation.setSearchMethod (pcl::search::Search<PointT>::Ptr (new pcl::search::KdTree<PointT>));
00107   vfh_estimation.setInputCloud (points);
00108   vfh_estimation.setInputNormals (normals);
00109   GlobalDescriptorsPtr global_descriptor (new GlobalDescriptors);
00110   vfh_estimation.compute (*global_descriptor);
00111 
00112   return (global_descriptor);
00113 }
00114 
00115 /* A simple structure for storing all of a cloud's features */
00116 struct ObjectFeatures
00117 {
00118   PointCloudPtr points;
00119   SurfaceNormalsPtr normals;
00120   PointCloudPtr keypoints;
00121   LocalDescriptorsPtr local_descriptors;
00122   GlobalDescriptorsPtr global_descriptor;
00123 };
00124 
00125 /* Estimate normals, detect keypoints, and compute local and global descriptors 
00126  * Return: An ObjectFeatures struct containing all the features
00127  */
00128 ObjectFeatures
00129 computeFeatures (const PointCloudPtr & input)
00130 {
00131   ObjectFeatures features;
00132   features.points = input;
00133   features.normals = estimateSurfaceNormals (input, 0.05);
00134   features.keypoints = detectKeypoints (input, features.normals, 0.005, 10, 8, 1.5);
00135   features.local_descriptors = computeLocalDescriptors (input, features.normals, features.keypoints, 0.1);
00136   features.global_descriptor = computeGlobalDescriptor (input, features.normals);
00137 
00138   return (features);
00139 }
00140 
00141 #endif


pcl
Author(s): Open Perception
autogenerated on Mon Oct 6 2014 03:15:05