detection of temporary objects in mobile lidar point clouds
TRANSCRIPT
![Page 1: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/1.jpg)
Detection of Temporary Objects in Mobile
Lidar Point Clouds
Xinwei Fang
Guorui Li
Kourosh Khoshelham
Sander Oude Elberink
![Page 2: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/2.jpg)
BACKGROUND
Mobile laser scanning provides an accurate recording of road
environments useful for many applications;
Usually only permanent objects are of interest;
Temporary objects can hamper the analysis of permanent ones;
Example: change detection using multi-epoch data where
temporary objects appear as (false) change signals.
2
![Page 3: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/3.jpg)
APPROACH
Static
Temporary objects:
Moving
Static temporary objects:
segmentation + classification based on shape features
Moving temporary objects:
segmentation + closest point analysis in two sensor data
3
![Page 4: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/4.jpg)
STATIC TEMPORARY OBJECTS
Classification based on shape features:
Linear
& short Linear
& tall
Volumetric
& tall
Planar Volumetric
& short
4
![Page 5: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/5.jpg)
MOVING TEMPORARY OBJECTS
Detection based on closest point analysis with two
sensor data
Sensor 1
Sensor 2
5
![Page 6: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/6.jpg)
METHODS
1. Ground removal
2. Connected-component segmentation
3. Static temporary objects:
feature extraction + classification
4. Moving temporary objects:
closest point analysis + classification
6
![Page 7: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/7.jpg)
GROUND REMOVAL
Plane-based segmentation
Ground = large & low segment
7
![Page 8: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/8.jpg)
CONNECTED COMPONENT SEGMENTATION
Points that are closer than a certain distance (e.g. 30 cm) belong
to one connected component;
8
![Page 9: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/9.jpg)
FEATURE EXTRACTION
Shape features
Size (number of points)
Area on horizontal plane
Mean density
Height of the lowest point
Height
Contextual feature
Distance to trajectory
Eigen-based features
Anisotropy
Planarity
Sphericity
Linearity
9
![Page 10: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/10.jpg)
CLASSIFICATION
Ground truth for training and evaluation: manual labeling (115 samples)
Feature selection
Forward Selection (FS)
Backward Elimination (BE)
Classification
Linear Discrimnant Classifier (LDC)
Support Vector Machine (SVM)
10
![Page 11: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/11.jpg)
CLOSEST POINT ANALYSIS
Static Moving
Static
objects
Moving
objects
Number 45 24
11
![Page 12: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/12.jpg)
EXPERIMENTS
Study area
Enschede, urban
Data
TopScan MLS
One strip: ~20 mio points
12
![Page 13: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/13.jpg)
RESULTS
Classifier Completeness Correctness
LDC - all features 0.90 0.86
LDC - FS (8 features) 0.90 0.79
LDC - BE (8 features) 0.95 0.91
SVM - all features 1.00 0.91
SVM - FS (9 features) 1.00 1.00
SVM - BE (11 features) 1.00 0.95
Detection results for static temporary objects (test set):
13
![Page 14: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/14.jpg)
RESULTS
Classification results for the whole strip :
14
![Page 15: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/15.jpg)
RESULTS
Sensor 1 Sensor 2 Sensor1 + Sensor2
Completeness 0.93 0.86 0.90
Correctness 0.90 0.96 0.93
Overall Accuracy 0.84 0.83 0.84
Detection results for moving objects:
Total 64 samples
4 false positives
6 false negatives
15
![Page 16: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/16.jpg)
LIMITATIONS
Occlusion;
Overgrown and undergrown segments;
Shrunk or elongated shapes due to
movement.
Variable shape and size of vehicles.
16
![Page 17: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/17.jpg)
SUMMARY
Object-based approach: obvious choice as we are dealing with
objects not points;
Connected component segmentation of objects works well (but
not perfect!);
Shape features are suitable for classification of static temporary
objects; Accuracy > 90%.
Distance between closest points is a useful measure for detecting
moving objects; Accuracy > 80%
Occlusion: objects seen by one sensor but not the other =
problem for moving objects.
17
![Page 18: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/18.jpg)
THANK YOU!
18
![Page 19: Detection of temporary objects in mobile lidar point clouds](https://reader033.vdocuments.us/reader033/viewer/2022052602/559afc3b1a28aba01a8b45ef/html5/thumbnails/19.jpg)
19