Given that about 95% of pictures get plate solved correctly, paying double or triple for the hosting service just to try to get the extra 5% solved didn’t make much sense, but I’m planning to migrate to another cheaper hosting to see if we can improve that number. I need to upgrade the system to have more memory but the current hosting is quite expensive. Some times the jobs time out before trying all index files due to memory constraints. I would really love to have the same index files because I’m not sure if the reason they are failing is because of the settings, the index files, or… hardware. When I noticed some images were being plate solved on Nova but not on Telescopius, I started trying different settings but none of them would work in 100% of the cases. I estimate over 95% of all deep sky pictures get correctly plate solved on Telescopius. ![]() They offer 2 or 3 different sets of index files, I’ve tried them all and none of them solves difficult pictures the same way with the default settings. ![]() ![]() Nova uses a set of index files unavailable to the general public due to licensing issues (they are called the “200-series”), while our server uses a different set of index files which naturally gives different results. Plate solving works by using index files containing the position of stars. There are many things going on, let me explainĪt the beginning Telescopius used the API to request plate solving of all pictures, but some shortages on their service, some delays in many cases due to job queues, and their suggestion to run plate solving on our own servers to avoid waiting for their global queue, made me switch the implementation to have it running on our own server.
0 Comments
Leave a Reply. |