update README

YuePanEdward · Feb 27, 2020 · 1711923 · 1711923
1 parent f95394d
commit 1711923
Show file tree

Hide file tree

Showing 6 changed files with 703 additions and 18 deletions.
diff --git a/LICENSE b/LICENSE
diff --git a/README.md b/README.md
@@ -11,11 +11,21 @@ The earlier conference version of GH-ICP is called Iterative Global Similarity P
 
 To highlight two key innovative points of the algorithm, we renamed IGSP as GH-ICP.
 
- ![alt text](img/gh-icp.gif)
+### Demo
+
+##### ETH TLS Dataset
+ ![alt text](img/demo_2.gif)
+
+##### WHU TLS Dataset
+ ![alt text](img/demo_1.gif)
+
+##### More
+ ![alt text](img/showresult2.jpg)
+
+### Principle
 
  ![alt text](img/GH-ICPworkflow.jpg)
 
- ![alt text](img/showresult2.jpg)
 
 ### [Paper](https://ieeexplore.ieee.org/abstract/document/8490968) 
 
@@ -28,11 +38,11 @@ If you find our work useful in your research, please consider citing:
           year={2018}
         }
 
-### original version available on Windows.
+#### original version available on Windows.
 Compiled with Visual Studio 12 2013 Win64 Release / Debug Passed, see [former release](https://github.com/YuePanEdward/GH-ICP/releases)
 
 
-### under code refactoring
+#### now available on Linux (passed on ubuntu 16.04)
 
 ### How to use 
 
@@ -66,9 +76,9 @@ sh script/run.sh
 using_feature=B;              # Feature selection [ B: BSC, F: FPFH, R: RoPS, N: register without feature ]
 corres_estimation_method=K;   # Correspondence estimation by [ K: Bipartite graph min weight match using KM, N: Nearest Neighbor, R: Reciprocal NN ]
 
-downsample_resolution=0.2;    # Raw data downsampling voxel size, just keep one point in the voxel  
-neighborhood_radius=0.6;      # Curvature estimation / feature encoding radius
-curvature_non_max_radius=1.8; # Keypoint extraction based on curvature: non max suppression radius 
+downsample_resolution=0.1;    # Raw data downsampling voxel size, just keep one point in the voxel  
+neighborhood_radius=0.5;      # Curvature estimation / feature encoding radius
+curvature_non_max_radius=1.5; # Keypoint extraction based on curvature: non max suppression radius 
 weight_adjustment_ratio=1.1;  # Weight would be adjusted if the IoU between expected value and calculated value is beyond this value
 weight_adjustment_step=0.1;   # Weight adjustment for one iteration
 registration_dof=6;           # Degree of freedom of the transformation [ 4: TLS with leveling, 6: arbitary ]

diff --git a/img/gh-icp.gif → img/demo_1.gif b/img/gh-icp.gif → img/demo_1.gif
diff --git a/img/demo_2.gif b/img/demo_2.gif
diff --git a/include/cloud_viewer.hpp b/include/cloud_viewer.hpp
@@ -39,7 +39,7 @@ class CloudViewer
     ~CloudViewer(){};
 
     void Dispaly2Cloud(const typename pcl::PointCloud<PointT>::Ptr &Cloud1, const typename pcl::PointCloud<PointT>::Ptr &Cloud2,
-                          std::string displayname, int display_downsample_ratio)
+                       std::string displayname, int display_downsample_ratio)
     {
         boost::shared_ptr<pcl::visualization::PCLVisualizer> viewer(new pcl::visualization::PCLVisualizer(displayname));
 
@@ -170,8 +170,9 @@ class CloudViewer
         }
 
         bool intensity_available = pcl::traits::has_field<PointT, pcl::fields::intensity>::value;
-        if (Cloud_T->points[0].intensity<0.001) intensity_available=0;
-
+        if (Cloud_T->points[0].intensity < 0.001)
+            intensity_available = 0;
+
         char t[256];
         std::string s;
         int n = 0;
@@ -190,7 +191,7 @@ class CloudViewer
                 pt.z = Cloud_T->points[i].z;
 
                 if (intensity_available)
-                    intensity_ratio = Cloud_T->points[i].intensity / 255.0;
+                    intensity_ratio = 1.0 * Cloud_T->points[i].intensity;
                 else
                     intensity_ratio = 1.0;
 
@@ -213,10 +214,10 @@ class CloudViewer
                 pt.z = Cloud_S->points[i].z;
 
                 if (intensity_available)
-                    intensity_ratio = Cloud_S->points[i].intensity / 255.0;
+                    intensity_ratio = 1.0 * Cloud_S->points[i].intensity;
                 else
                     intensity_ratio = 1.0;
-                
+
                 pt.r = 255 * intensity_ratio;
                 pt.g = 215 * intensity_ratio;
                 pt.b = 0 * intensity_ratio;

diff --git a/script/run.sh b/script/run.sh
@@ -5,15 +5,15 @@ output_point_cloud_path=/media/edward/BackupPlus/Data/ETH_registration_TLS/arch/
 
 #parameters
 using_feature=B;              # Feature selection [ B: BSC, F: FPFH, R: RoPS, N: register without feature ]
-corres_estimation_method=K;   # Correspondence estimation by [ K: Bipartite graph min weight match using KM, N: Nearest Neighbor, R: Reciprocal NN ]
+corres_estimation_method=N;   # Correspondence estimation by [ K: Bipartite graph min weight match using KM, N: Nearest Neighbor, R: Reciprocal NN ]
 
-downsample_resolution=0.2;    # Raw data downsampling voxel size, just keep one point in the voxel  
-neighborhood_radius=0.6;      # Curvature estimation / feature encoding radius
-curvature_non_max_radius=1.8; # Keypoint extraction based on curvature: non max suppression radius 
+downsample_resolution=0.1;    # Raw data downsampling voxel size, just keep one point in the voxel  
+neighborhood_radius=0.5;      # Curvature estimation / feature encoding radius
+curvature_non_max_radius=1.0; # Keypoint extraction based on curvature: non max suppression radius 
 weight_adjustment_ratio=1.1;  # Weight would be adjusted if the IoU between expected value and calculated value is beyond this value
 weight_adjustment_step=0.1;   # Weight adjustment for one iteration
 registration_dof=6;           # Degree of freedom of the transformation [ 4: TLS with leveling, 6: arbitary ]
-appro_overlap_ratio=0.6;      # Estimated approximate overlapping ratio of two point cloud 
+appro_overlap_ratio=0.5;      # Estimated approximate overlapping ratio of two point cloud 
 
 launch_realtime_viewer=1;     # Launch the realtime registration viewer during registration or not (1: Launch, 0: Not launch)