Drive error analysis


A window like the following appears at the end of the batch astro-process to show how images have moved during the sequence that has been processed. If the camera was on a fixed tripod the apparent movements will be much larger, due to the rotation of the earth. The example used a motor-driven equatorial mount and shows how the drive (inevitably) is not perfect.

In this display, dx and dy are the average displacements horizontally and vertically between successive images, in pixels (px). The dr column shows the root mean square of dx and dy, ie, the total displacement in the Pythagorean sense. All of these measurements are effectively between the middle of one exposure and the middle of the next, so the length of each exposure is also tabulated to enable the gap between exposures to be eliminated when estimating whether there was a large movement during one exposure. You could use this to decide not to include certain images in the processing, and re-run it.

The image size from my camera is 5616 x 3744 pixels, so a 7-pixel movement is a very small percentage error.