Workshop "Automatic Methods for Object Detection from Very High Resolution Satellite Images"
lecturer
- Dr. Beril Sirmacek, Deutsches Zentrum für Luft- und Raumfahrt (German Aerospace Center), Cologne, Germany
date
- 10:00 - 12:00 and 13:15 - 15:00 on May 19, 2011
- 13:00 - 15:30 on May 20, 2011
location
- Z 613, University of Konstanz, Konstanz, Germany
lecture part 1
from 10:00 to 12:00 on May 19Introduction to remote sensing:
Introduction, History and scope of remote sensing, Electromagnetic radiation and its properties, Sources of remotely sensed data: active sensors (optical, near infrared, thermal Images), Passive Sensors (SAR, LIDAR data) remote sensing applications in real life.
Characteristics of Remotely Sensed Images:
Pixel term, Spatial resolution, Spectral resolution, Illumination and view angle effects.
Remotely Sensed Image Analysis:
Pre-processing (noise reduction, rectification), Image Enhancement Techniques (histogram enhancement, histogram equalization, gamma correction, thresholding), Edge detection, Image processing in Frequency domain (2D Fourier transform, and smoothing, sharpening, noise filtering in frequency domain).
lecture part 2
from 13:15 to 15:30 on May 19Remotely Sensed Image Analysis (cont.):
Image segmentation, Local feature extraction (SIFT, Harris), Geometrical features (compactness, rectangularity).
MatLab Examples:
- Noise reduction techniques
- RGB color bands of an aerial image, forest/road/building detection examples
- Texture detection using gray-level co-occurance method
lecture part 3
from 13:00 to 15:30 on May 20Advanced Methods for Object Detection from Remotely Sensed Images:
Challenges in object detection, Short overview of object detection studies in satellite images, Object detection using templates, Object detection without using templates, Texture detection, Change detection.
Applications of Remote Sensing:
Automatic land mapping using object detection techniques, Three-dimensional surface model generation (methods for obtaining digital surface models from LIDAR data, airborne images, and satellite images), Three-dimensional city model reconstruction, Other applications of remote sensing: examples on using remote sensing and automatic object detection techniques in archaeological studies.
MatLab Examples:
- Automatic object detection in satellite images using SIFT features
- 3D building reconstruction example
- Detection of an archaeological region from panchromatic satellite image automatically.