Master Thesis name: Extraction and 3D characterization of green features in urban areas
from Pléiades image
Supervisor: Assoc.-Prof. Dr. Stefan Lang
Keywords:
Pléiades,
Support Vector Machine, nDSM, 3D green feature, urban vegetation
Used software for implementation:
Software
|
Version
|
Usage
|
Envi
|
IDL 8.3
|
Radiometric Calibration - TOA Reflectance
|
Erdas Imagine
|
2016
|
Orthorectification and Definition of Projection
|
eCognition
|
Developer 9
|
Image processing, segmentation, classification, accuracy assessment
|
ArcMap
|
ArcGIS 10.4
|
Image clipping, projection
|
ArcScene
|
ArcGIS 10.4
|
3D feature visualization and analyses
|
Methodology:
DIGITAL SURFACE MODEL GENERATION METHOD
DSM OUTLIERS CLEANING
OBJECT BASED IMAGE ANALYSIS METHODOLOGY FOR VEGETATION
VEGETATION CLASSIFICATION
Spectral signature
Support Vector Machine algorithm
Combination of visual and digital interpretation for shadow removing
ACCURASY ASSESSMENT
The importance of accuracy assessment and the matter of accuracy
error
Required considerations before accuracy assessment
Sampling method
Accuracy assessment of classification results
Field measurement and visual investigation
ANALYSES ON ARCGIS
Mapping of third dimension of vegetation with ArcGIS
Characterization of vegetation on ArcGIS
The capabilities and limitations of tree volume estimation
OBIA approach for tree crown delineation
Analyses with tree crowns and difference
