Tree branches confuse astroalign
Short-bus opened this issue · comments
The attached images show a case where there are trees in the field of view and it seems to confuse the astroalign routine.
The branches are being recognised as potential stars.
As a workaround have suggested trying shorter exposure times in TrackingExposureSeconds, reducing from 5 seconds to 1 or 2 seconds to see if a darker image helps the real stars stand out better in the image.
Original Tracking_Analysis image
Original Latest_Tracking image
Original Target_Tracking image
Adding experimental new parameter "TrackingUrbanFilter".
Default (False) - means no effect.
When set to True, the program uses OpenCV to try to remove 'haze' from the image.
It also reduces the impact of tree branches in the image.
The filter blurs the image horizontally to establish a background light level for the image.
This blurred image is subtracted from the live captured image.
The result is then passed to the pilomarimage.EnhanceStars() method.
EnhanceStars() applies OpenCV's Gaussian Blur and Threshold methods to isolate the stars that remain.
Needs to be tested and tuned with more observations.
Before applying the UrbanFilter:
After applying the UrbanFilter:
2024-01-issues has a workaround in the new filter scripts, these can be further adjusted now via the parameter file.