Relastro @ ITP - Goethe University, Frankfurt | Clustering and averaging the images of Sgr A* @relastroitp-goetheuniversi1616 | Uploaded 2 years ago | Updated 2 hours ago
This animation starts by showing a time-lapse video of the Dolomites mountain range, illustrating the process of image clustering and averaging used to image the Milky Way's central black hole, Sagittarius A*. The video showcases why, in a long-exposure observation of a variable subject, it is possible to recover multiple possible images of the same mountain range. The various images produced are sorted into four different categories - known as clusters - according to their main features. Each cluster is accompanied by a vertical bar that indicates how often an image of that cluster is recovered from the total set of images. The images in each cluster are then averaged in the bottom panels and a final average image is constructed in the top part as a a weighted average of the various cluster averages (a cluster with a higher vertical bar has more weight in the final average image). The second part of the video then shows this process applied to the actual images of Sagittarius A* recovered from the Event Horizon Telescope observations of the black hole. In this latter case, the different images represent equally good fits of the observational data and do not refer to different instants in time as in the time-lapse movie.
Credit:
C. M. Fromm (University Würzburg, Germany), L. Rezzolla (University Frankfurt, Germany), EHT Collaboration
Link to our homepage: https://relastro.uni-frankfurt.de/
Link to JETSET: jetset-erc.org
Link to the EHT website: eventhorizontelescope.org
Link to the Black Hole Cam Website: blackholecam.org
Link to the Relativistic Jets in Active Galaxies website: http://www.for5195.uni-wuerzburg.de/
This animation starts by showing a time-lapse video of the Dolomites mountain range, illustrating the process of image clustering and averaging used to image the Milky Way's central black hole, Sagittarius A*. The video showcases why, in a long-exposure observation of a variable subject, it is possible to recover multiple possible images of the same mountain range. The various images produced are sorted into four different categories - known as clusters - according to their main features. Each cluster is accompanied by a vertical bar that indicates how often an image of that cluster is recovered from the total set of images. The images in each cluster are then averaged in the bottom panels and a final average image is constructed in the top part as a a weighted average of the various cluster averages (a cluster with a higher vertical bar has more weight in the final average image). The second part of the video then shows this process applied to the actual images of Sagittarius A* recovered from the Event Horizon Telescope observations of the black hole. In this latter case, the different images represent equally good fits of the observational data and do not refer to different instants in time as in the time-lapse movie.
Credit:
C. M. Fromm (University Würzburg, Germany), L. Rezzolla (University Frankfurt, Germany), EHT Collaboration
Link to our homepage: https://relastro.uni-frankfurt.de/
Link to JETSET: jetset-erc.org
Link to the EHT website: eventhorizontelescope.org
Link to the Black Hole Cam Website: blackholecam.org
Link to the Relativistic Jets in Active Galaxies website: http://www.for5195.uni-wuerzburg.de/
