enpeda Research Group

Auckland University of Technology

Professor Reinhard Klette

Prof. Reinhard Klette (Auckland University of Technology, Fellow of the Royal Society of New Zealand)
made significant contributions to two major areas, digital geometry and computer vision. Dr. Reinhard
Klette co-authored (together with the late Prof. Azriel Rosenfeld of University of Maryland, USA – the
most prominent scientist in this area) a book on Digital Geometry, published in 2004 by Morgan
Kaufman San Francisco, which defined the field and has been cited more than 500 times.
Dr Klette has been working in the area of computer vision for more than 30 years. He has become
internationally renouned for his work in vision-based driver assistance since 2006, with important
contributions on performance evaluation and improvements of correspondence algorithms (for stereo
matching and optical flow) on real-world video data, supporting, for example, 3D scene reconstruction
from a mobile platform.

In 2008 he co-authored (with two of his former PhD students) a research monograph on panoramic
vision (with Wiley, UK), and in 2011 a research monograph (also co-authered with a former PhD
student) on shortest paths in Euclidean spaces (with Springer, UK).

The enpeda project is basically about a special subject in computer vision: how to understand or model a 3D environment based on multiple image sequences, recorded by `normal' or specialized (such as night vision) cameras.
However, it also addresses visualization of 3D environments, using recorded stereo sequences or computer graphics for synthesizing 3D scenes or geometry, and human-machine interaction to some extent. This defines the project in the wider area of multimedia imaging. The main application scenario is as follows: cameras are installed in a vehicle (typically a car, but possibly also a wheelchair, a forklift, a boat, and so forth), and the operation of this vehicle (by a driver) should be supported by analyzing video sequences recorded by those cameras. Possibly, further sensor data (e.g., GPS, radar) are also available to be analyzed in an integrated system. The image of a New Zealand weta (on the right) is the project's logo. The use of two eyes and of an internal model about the 3D environment is also the basic mechanism for this insect to operate in complex, often unpredictable environments. The project was possible due to initial support (in 2006 and 2007) by a partner at Daimler AG, Germany. Due to further sponsors the project has access to a test vehicle (a 2007 Mercedes car, called HAKA1) with proper access to the data recorded by a modern car computer. Lectures on Multimedia Imaging provide basic knowledge, enabeling students to contribute to this project. Students will capture their own stereo sequences in HAKA1. This process actually comprises camera calibration, recording of stereo videos, sequence rectification, and sequence uploading into the projects sequence data base.

rklette@aut.ac.nz