MSI-RCT (3.0) tilt tracking problem....
Added by Duncan Sousa about 7 years ago
Hi,
We're having problems with getting MSI-RCT (3.0) to work. I think it's a tracking issue, but maybe that is a downstream problem caused by some other setting. RCT seems to find good peaks when taking the test images at different tilts using the RCT node "acquire" and "tilt stage alpha..." buttons. It also seems to start off tracking okay, but fails and repeats the same angle 15 times (getting progressively worse each time) until it gives up. I've attached a few images with the messages and settings for the last settings I tried (although I've tried a large number of various settings at this point from different angles to filter settings. I noticed a couple other oddities. It seems to ignore the maximum tilt stepsize setting. Once it starts to fail it also seems to progressively threshhold the tilt tracking images in such a manner that there are no features left to view in a binary black and white image.
Thanks,
Duncan
| 20170825_173831.jpg (3.22 MB) 20170825_173831.jpg | |||
| 20170825_173637.jpg (5.81 MB) 20170825_173637.jpg | |||
| 20170825_173834.jpg (6.26 MB) 20170825_173834.jpg | |||
| 20170825_173846.jpg (8.28 MB) 20170825_173846.jpg | |||
| 20170825_173854.jpg (4.65 MB) 20170825_173854.jpg |
Replies (3)
RE: MSI-RCT (3.0) tilt tracking problem.... - Added by Anchi Cheng about 7 years ago
Hi, Duncan,
Your image has a tear which makes geometrical assumption of what it looks like when tilted not predictable with a simple model of a flat plane which is RCT Sift algorithm uses. I think that is certainly a good reason.
If you have a case in which you don't have floppy bits and it still fails, could you attach a pair of images it need to match, i.e., the angle where it fails and the tilt angle before that, not as screenshot but the real images in either jpg or mrc form. I can then run the algorithm and see if I can get it to work.
RE: MSI-RCT (3.0) tilt tracking problem.... - Added by Duncan Sousa about 7 years ago
Hi Anchi,
I was afraid the problem was something like that. I'll post images when I next get a chance for RCT testing. Unfortunately, nearly every square on every grid we image is warped and not a flat plane. However, with the RCT Sift model in mind I'll see if I can get things to work and what the limits are if this turns out to be the issue. Maybe using a much higher mag where the carbon is locally flat would help (e.g. 2500x).
Duncan
RE: MSI-RCT (3.0) tilt tracking problem.... - Added by Anchi Cheng about 7 years ago
Yes, higher mag would help in this particular case to concentrate on the flat region.