Hao Zhou did his honours project at the Chair for Computation in Engineering (CIE) at the TU München.
Feature extraction is one of the most important image processing techniques. By low-level feature extraction, features can be extracted automatically from an image without any shape information. By high-level feature extraction, information such as the shape, size and position of the object in image can be obtained.
In this project we developed a geometry extraction tool to automatically extract the outlines of buildings from the digital image and generate a XML file which stores the geometry of these buildings.
The implementation contains various image processing and feature extraction techniques such as Canny edge detection, Hough transformation, and polygon detection. All techniques were implemented in Java.
The 2D Geometric Objects that are extracted by the geometry extraction tool are used as obstacles in a pedestrian simulation system.
The two figures come from the report.