This has been my biggest issue (bug) with PI. Happens a great deal for me.
Same symptoms as reported: RANSAC goes thru tries as high as approx #20 (at which point it has been 5 or 10 minutes on the same image) and I must crash the program to get it to quit. I have let it go for as long as 2 hours. It will never quit. Some images go as high as #3 and work fine. Seems if it gets to #4 (which only takes it a few seconds, it then continues to cycle and never quits. Each cycle runs much slower than the prior cycle. If it gets past about #12 it won't even respond to the Pause/Abort button and I have to crash the program.
It takes about 3 minutes to get to #16 on my pc, then gets dramatically slower. At 1 to 2 hours it has only gotten to about #21.
My pc: i7-3930, 4.4 ghz, 32gb memory. (that's 12 processors running at 4.4 ghz, and they are all running, often at %100). So this process is clearly never going to quit.
What triggers it? Images that have a significant star trail, on the order of > 2x.
I have tried setting "Maximum Stars" to values between 500 and 10000. This produces an error then continues with the next image, even for images that do work when it is set to Auto.
It seems to me this routine badly needs some logic that says... quit after 5 cycles. "RANSAC Iterations" can only be set to a minimum of 20. Not sure if this is the parameter that control the number of tries.
Could this be a good place to add a feature to automatically reject images that do not meet a minimum quality standard?
