First of all, thank you for your consideration.
My idea of implementation would be something like this:
1) Extend the exposure time of the wide-angle lense/sensor group to at least a second and if possible make it adjustable by the user.
2) Give the option to the user (if the telescope is in EQ Mode) to select a star in the wide-field image and "pin" it as the guide star.
3) Write the code to allow the onboard computer of the Seestar S30 to detect motions of this guide-star in the wide-field image (for instance of a few pixels) and calculate the movement of BOTH axis of the S30 (in this case we can call them RA and DEC) necessary to compensate for that motion.
4) Last but surely not least: when this new autoguide feature is active, leave the user the faculty to extend integration time per frame as much as they like, or at least up to a 2-3 minutes.
Of course point 4 would imply that darks of the same exposure should be also taken.
Surely the difference of focal length between the two group/sensor groups will not allow for very precise tracking and corrections, however I think it should be worth the try.
Best Regards
