Máscara para detecção de detritos espaciais em imagens de telescópio adquiridas em modo estático

Abstract

Orbital debris approximately 10 cm in size and larger can be monitored with ground-based telescopes and radar. These debris threaten the operation of satellites and impact the economy and global security of space activities. In the geoestatiory orbit (GEO), where most of the highest economic value satellites are located, there are approximately 842 cataloged debris. In the Low Earth Orbit (LEO) there are approximately 13485 cataloged debris. However, European Space Agency (ESA) studies show that hundreds of millions of small objects above 1 mm are currently in the two GEO and LEO orbits above Earth and have not yet been catalogued. In this work we created a computational procedure to detect possible space debris in GEO orbits with images obtained from telescopes on land and in Tracking Rate Mode, where the stars in the sky background appear in the form of lines in the CCD images and the garbage in the form of points. CCD images of 2092 x 2092 pixels (high resolution), with 5 degrees of field of view (FOV) and 7 seconds of exposure, used in this work, were obtained with the PanEOS telescope (Panoramic Electro-Optical System), 750 mm of opening, installed in the observatory of Picos dos Dias of the National Laboratory of Astrophysics (LNA). For this research we adapted the Photutils packages written in Python to build a mask and separate stars from candidate space debris. Our methodology consisted of first smoothing the images using a Gaussian filter, then each element was tagged in different categories and finally the stars were erased, resulting in only the space debris candidates. We test flux combinations to establish the detection limit and use different points spread function (PSF) to determine the elongation limit of objects. Our methodology works with a single image at a time quickly and efficiently and allows detecting objects with different PSF and thus requires low hardware capacity. Our results in this validation phase identified 76% of the artificial training debris and in the real images of the PanEOS telescope we detected real objects and consistent with a possible space debris. Finally, it is concluded that the algorithm allows the reading of a database of real images like the ones we have from the PanEOS telescope and is the first step to catalog space debris and find the size.