6 questions with Udo Tempelmann, Leica Geosystems
24 January 2012
Udo Tempelmann, Manager System Engineering at Leica Geosystems gives us the inside line on the type of imaging technology behind global surface imaging and mapping systems like Google Earth...
Q: Please briefly describe your background in digital imaging.
A: After studying "Geodesy and Surveying" with a diploma thesis about image processing, I spent 6 years at the Institute for Photogrammetry and Image Understanding of the University Bonn, Germany. Besides working in remote sensing projects, I got involved in the MOMS (modular optoelectronic multispectral stereo scanner) project on data processing side.
MOMS was the first three-line-camera, a multi pushbroom sensor with three viewing directions (forward, nadir, backward). That principle allows a reconstruction of the trajectory to the high requirements of photogrammetry. Shortly after MOMS was flown in the Space Shuttle, I joined Leica as a photogrammetric software developer. It took, however, only a few days to get involved in the design of a new scanner for aerial photos.
A few years later, the three-line-camera came unexpectedly back upon me, when Leica started in 1997 in close cooperation with the DLR (German Aerospace Center) the development of the ADS (Airborne Digital Sensor) multi-line pushbroom sensor. Since then, I have been active in the aerial sensor development, as well on system level, as on many details in the fields of optics, mechanics, thermal management, CCD sensors, calibration and software development.
Q: Please can you define photogrammetry and orthophoto?
A: In simple terms, photogrammetry is the art of geometric measurement (the "metry") in photograms (today called photographs). This is especially the 3D-measurement points and higher features from stereo-pairs.
An orthophoto (ortho = right, true) is the result of a projection from the view of a camera (classically a central perspective view, for pushbroom a combination of parallel perspective along and central perspective across the image strip) to a vertical parallel (an "orthogonal") projection. Mostly this is used to fit a mosaic of orthophotos to a mapping system. Necessary for the creation are images with known orientation and a surface model onto which the images are projected.
Just have a look at Google Maps or any similar image map service to see a plethora of orthophotos.
Q: What has changed in this field in the last 5 years?
A: The past five years were really the time where digital aerial imaging replaced the photography on film. The switch to digital came later than in other fields, as the very high resolution requirement (standard photo size on film was 23cmx23cm) could not be easily fulfilled by digital sensors. Meanwhile frame cameras of up to 252 megapixels or pushbroom scanners that can deliver image strips of many gigapixels are in use.
Q: What are the main drivers for change and technology development?
A: The world, including photogrammetry, has become digital, so it was logical to avoid the cumbersome steps of processing and scanning large format films.. also the applications of aerial photogrammetry have changed. Manual map extraction from stereo pairs, the classical photogrammetric task, is still done but automatic extraction of surface models, features like roads or buildings, and especially the production of ortho photos are gaining more and more importance.
Digital sensors can also capture a much higher information depth than panchromatic film, colour diapositive or false colour diapositives, that were common before. A typical digital sensor can simultaneously capture at least the four important spectral bands of visible (RGB) and near infrared with a far better radiometric resolution and well-defined spectral sensitivity.
Q: Can you outline the main imaging challenges for you and your business at present?
A: What pushes us into new developments is now mainly the economic aspect. Besides improved area performance (measured in gigapixels per hour of flight), also higher sensitivity is important, as it extends the number of flight hours per day. Last, not least, the reliability of the system and the usability of images that are taken under uncommon conditions (under clouds or with reflexes on water and man made objects) need our attention. Current sensors deliver reliably about a Terapixel per flight day, but that number is constantly growing.
Q: Finally, we are pleased to have you on board for the conference this year, what are you hoping to gain from the conference?
A: Being just a single-purpose user of image sensor technology, I hope to generally broaden my view of what's going on in the sensor development, how sensors are used in other fields, and what we can expect for the future, regarding sensor size, performance and spectral sensitivity.
Udo Tempelmann will be delivering a presentation at Image Sensors 2012 entitled 'Image sensors for photogrammetry and orthophoto production'