59th International Astronautical Congress
Glasgow, Scotland, 29 September - 3 October 2008
Automatic GEO Object Detection Algorithms and their Implementation in the Framework of the Apex II Image Processing System
Vladimir Kouprianov
Central (Pulkovo) Astronomical Observatory of the Russian Academy of Sciences Saint Petersburg, Russia
Large Field of View Survey Cameras
- Efficient, but
- Produce large amount of data (both orbital objects and field stars)
Typical Large-FOV Image
- Overcrowded with star trails
- Several dozens GEO objects are hard to detect
- Need to obtain maximum information from a single image
|
|
Optical Aberrations
|
|
Star trail (image center) |
Star trail (image corner) |
Apex II Software Package for Image Processing
- General-purpose open-source astronomical image processing platform
- Modern technologies – Python (flexible scripting programming language), packages for scientific computations and graphics, easily extendable modular structure
- Used by more than 20 teams participating in and cooperating with ISON
Residual Image After Star Trail Subtraction
- Complete elimination of star trails would take a lot of CPU time
- Impossible due to distortion of trails by optics and atmosphere
|
|
Object Detection: Binary Image
- Residual image still contains much more star trails than GEO objects; how to eliminate them?
|
|
Binary Image Filtering for Star Trail Elimination
Filtered Binary Image
- Almost no star trail remains left
- Remaining spurious detections are easily removed by comparing adjacent frames
|
|
Conclusions
Automatic processing of images taken by wide-field survey cameras require efficient algorithms for star trail elimination
Such an algorithm, based on logical filtering of a binary image, is proposed and implemented in the framework of the Apex II image processing system
2 октября 2008
Материалы докладов публикуются с согласия авторов
|