The 58th Session of International Astronautical Congress (IAC)
September 24-28, 2007
Hyderabad, India
 |
FIRST ITALIAN SPACE DEBRIS OBSERVATORY:
THE IMAGE PROCESSING AUTOMATION
F. Paolillo, M. Porfilio, F. Piergentili
Gruppo di Astrodinamica dell'Universita degli Studi "La Sapienza" - Agenzia Spaziale Italiana |
 |
OBJECT
Developing a software for the automatic detection and identification of Earth orbiting objects on frames from optical observations
Raw frame taken during an observation campaign (in 2002) heavily affected by noise and noise gradient.
SINGLE FRAME PROCESSING
|
- Each original image is divided in a number n (user defined) of sub-array.
- Evaluating the median values at the edges of the sub-array it is possible to estimate two noise gradient sub-matrices in vertical and horizontal directions.
- A flat noise sub-array is achieved by subtracting the gradient sub-arrays from the original frame.
- The overall “clean” image is just the composition of the sub-matrices.
|
TWO-FRAME PROCESSING
|
- To each frame is subtracted the previous one in order to delete most of the luminous pixels belonging to stars and biases.
- The algorithm is able to remove spurious pixels too.
 |
OBJECTS DETECTION
The histograms show on x axis the distances (in pixels) among all the couples of “luminous” pixels in the frame; the y axis reports the number of occurrences for each distance. The number of peaks is related to the number of satellites trails in the image.
 |
 |
| The histogram reports random distances because it belong to a frame without objects. |
- The first peak are the distances between the luminous pixels inside each individual object trail.
- The second one is due to the mutual distances between pixels belonging to the two objects
|
OBJECTS IDENTIFICATION
The measured angular positions of the detected objects are compared with the propagated positions of the bodies from TLE catalogue.
CONCLUSIONS
- A software for automatic image processing and objects identification has been developed and it is ready for operations;
- The software has been satisfactorily tested by means of “worst case” frames, affected by significant noise and noise gradient;
- A debris of magnitude 16.8 has been automatically detected by the software (the magnitude limit of the frames was about 17).
29 сентября 2007
Доклад публикуется с разрешения авторов
|
