extract_od is excessively slow with Nautilus data

[lenafb]

When analyzing images from the same plate (acquired on different machines) it takes multiSero ~7.5x longer to extract ODs from Nautilus images than Scienion images. Most likely this is due to the fact that multiSero looks for the well before applying masks and particle filter steps, hence WARNING:root:Couldn't find well in {current well} error.

To Reproduce
Steps to reproduce the behavior:
To best reproduce illustrate this behavior, I recommend running multiSero debug configs with the following commands (main difference being path to Scienion data vs path to Nautilus data):

Scienion run:
-e -i "/Volumes/GoogleDrive/Shared drives/compmicro/ELISAarrayReader/images_scienion/2021-08-20-12-09-52-head_to_head_plateII/E_to_H" -o "/Volumes/GoogleDrive/Shared drives/compmicro/ELISAarrayReader/images_scienion/2021-08-20-12-09-52-head_to_head_plateII/E_to_H" -d

Nautilus run:
-e -i "/Volumes/GoogleDrive/Shared drives/compmicro/ELISAarrayReader/images_nautilus/2021-08-20-VLP_RVP_NS1/E_to_H" -o "/Volumes/GoogleDrive/Shared drives/compmicro/ELISAarrayReader/images_nautilus/2021-08-20-VLP_RVP_NS1/E_to_H" -d

Expected behavior
multiSero should be able run extract_od in <5 min.

Operating environment (please complete the following information):

• OS: Mac, Jetson Nano

Additional context
The strategy I'll pursue is to suppress multiSero's effort to search for the well boundaries, since Nautilus data often does not include well boundaries.