Abstract: According to various embodiments, a telescope is automatically aligned without requiring user intervention and without requiring knowledge of actual local time or location. A mount model specifying a relationship between a telescope's internal coordinate system and a celestial coordinate system is generated using an arbitrary time, arbitrary telescope location, and a number of alignment reference points. A pointing error for the initial mount model is determined, for example using a plate solving technique to translate between plate coordinates and celestial coordinates for the alignment reference points. Time and location values are iteratively adjusted to reduce the pointing error until it is acceptably low. In one embodiment, adjustments are made by reference to a local sidereal time (LST) offset and/or a latitude value. In one embodiment, the iterative adjustment is performed using a two-phase methodology, including a coarse adjustment followed by a fine adjustment.

Abstract: A plate solving methodology determines celestial coordinates of an image. Star locations are extracted from the image in terms of pixel coordinates. A group of four stars, referred to as a “test quad”, is identified. A signature for the test quad is generated. In one embodiment, this test signature is derived by determining the separations of the four stars in the test quad, normalized by the largest separation. In one embodiment, the signature also includes the sum of these normalized separations. A query is performed, using the generated signature, against a database of reference signatures for known groups of stars (referred to as “reference quads”). A geometric transform is determined, establishing the relationship between the test quad and a reference quad that matches within a specified tolerance. This geometric transform defines the celestial coordinates of the image. Additional verification steps can be performed to confirm the accuracy of the match.