Add support for a water droplet detection and mitigation

[illigtr]

I designed my own AllSkyCam enclosure using ASA filament, a 3 inch poly carbonate dome and O-rings. So far, very pleased with the results, as I wish to have a unit permanently installed on the roof during all 4 seasons, with temperature variations of -30C to +40C, 15% RH to 100% RH.

DEW mitigation is simple enough already supported. Thank you!

Another issue is RAIN. During high humidity events, rain droplets stay on the poly carbonate dome for a very long time, hours...

I have been able to mitigate this somewhat, by treating the dome with a commercial product call RAIN-X. This does not cause distortion and reduces the size and number of droplets that stick to the dome surface.

I think mitigation could go further by installing a vibration device similar to mobile phones. The hardware is simple, any tiny DC motor with an intentionally off centre counterweight, driven by an NPN transistor connected to a GPIO.

The new feature to support this would:

1. Allow the vibrate device to run for a specific interval, something like .5 seconds to 3 seconds.
2. Trigger the vibration device between image captures IF humidity is greater than a specific value or perhaps more complicated, the detection of water droplets on the last image.

Love to help to make this a reality.

[illigtr]

After looking at all the options (so many! :) ) I think I may be able to use image pre/post save hooks to at least test the concept. I'll start there.

[aaronwmorris]

I have not really responded, but you are on the right track.

I am working on an enhancement to the hooks right now that will help you with this function.

[aaronwmorris]

I have merged #2239, but do not celebrate too soon.

After thinking about the hooks and where they are executed, there is no guarantee the camera is idle when these scripts are executed. indi-allsky is a multi-process/threaded application. Camera control and image capture is performed in the capture process, and image processing (where the hooks are run) is performed in the image process. These processes run independently and concurrently. In order to block the image capture, a hook would need to be added to the capture process. I am not opposed to adding this.

If you want to keep looking at the image hooks, there are example scripts in the ./misc/ folder.

[illigtr]

Ok. Understood. I can still test the concept by reading the various variables (external humidity, weather events, other) and then testing a short vibration burst to see if it clears the polycarbonate camera dome. Are there hooks in the camera control and image capture, or hooks planned?

[aaronwmorris]

The hooks are currently in the image process. I will work on a post-capture hook in the capture process.

[aaronwmorris]

I have merged #2245 which adds a pre-capture hook that runs in the Capture worker process. This hook is guaranteed to be run when the camera is idle.

[illigtr]

Thank you, I will run an update and then test. I have a prototype vibration device set up for testing.