Request for help: No extrema found in contour of xxx pixels in level xxx about py-eddy-tracker

[AntSimi]

Several things come to me, you need to apply a large scale filter to remove large scale structure like in this example.
Input grid have which resolution?

[cdmpbp123]

Thanks for your response and suggestion. I have tried different scale filters with 400, 700 and 1000km. The results with 1000km seems better than without filter, but most eddies detected in small region still can not be detected. And warning 'No extrema found in contour of xxx pixels in level xxx) still existed. Could you briefly explain why this warning happen? Is this because sla contour of my data so complicated that the algorithm can not find closed contour?
The input grid I used is 0.03°. How should high-pass filter value be determined, and could you provide a recommendation for my case?
Thanks a lot!

[AntSimi]

My first guess.
Contour detection algorithm could be high sensitive to high frequency data or noise. You could try to apply a low pass filter to smoth/simplify a little bit.

[cdmpbp123]

Thanks for your suggestion. I've tried several filters but still can't get satisfied results. I then divide the domain to several pieces and do the detection then concatenate them. The results seem OK now. Maybe the reason is the data quality, I will check the data later. Thanks for your help!

[AntSimi]

If you divide domain, result are different? i don't like this algotrithm behaviour., There is something to dig.

[cdmpbp123]

Yes. I test without any filter. Results vary significantly depending on the region dividing.