AXE045 Colour Sensor focussing

BDG

New Member
Can anyone tell me whether the AXE045 Colour Sensor can be focussed on infinity. i.e the sky. I was hoping to identify the condition of the sky - clear blue sky, light floffy cloudy, black and stormy cloudy etc. to correlate with the data I'm reading back from a Solar Panel array as well as provide some warning for impending (dangerous) storms.

I haven't played with one, but all the data seems to be related to near field identification of objects, not anything further away than a foot or so, and also the angle of view doesn't appear in the spec sheets I've read.

Thanks
 

Andrew Cowan

Senior Member
I have one, but I haven;t played with it yet.

The lens does unscrew, but I think it would just measure the colour of light, not the sky colour. An LDR would probably be all yo need to measure the sun getting through the clouds.

Good luck,

A
 

BeanieBots

Moderator
A very interesting idea.
I'm certainly not aware of any commercially available units which take into consideration sky colour but 'human' predictors certainly do so this should prove to be a very interesting project.

Does it actually NEED to focus on the sky. Surely, you just want an overall colour indication. However, if as Andrew says, the lens can be screwed in/out, then I'm sure infinite focusing would be possible. How you would know when it's focussed would be another question.

The more typical light level only sensors often have a semi opaque light diffuser fitted to deliberately AVOID actual focus so you might want to consider a similar setup for sky colour sensing. Also, bear in mind that direct sunlight if focused is very likely going to damage the sensor and even just a bright day is very likely to saturate the sensor so care needs to be taken to avoid these situations.

Before spending money on a sensor, it might be worth trying a couple of LDRs with coloured filters just to see if there is much discernable difference between the perceived sky colour variation and what can actually be measured.
 

BDG

New Member
Thanks for the ideas everyone.
Good point about direct sunlight damaging the sensor, i didn't think of that, maybe the LDRs with filters would be a good cheap quick check. I'd need to get the filter colours right thou to cover the spectum accurately. I would be concerned about non linearitues in the sensitivity of the LDRs though. I'd have to calibrate somehow.

As storm clouds can be quite localised. I wanted to be able to be sure that the sensor was reading the colour overhead, not just an average over the whole sky. Also directiion of the approaching clouds might be useful too. Dangerous hale storms often come from the south west, wind storms from the east, and rain from the North West. Actually detecting Lightning (brief pulses of light) could be detected too (as well as the obvious RF interference).
I'm sure flat LDRs would do that by virtue of it geometry, ie.e greatst at 90deg falling off as the sine of the angle.

Just thinking about the potential (mainly UV) sun damage (long term), an alternative might be to use light reflected off a white surface illuminated by the sky, maybe with a plastic cover like Hippy suggests to restrict the UV as well.

Some clever video processing might be the answer for someone with lots of DSP knowledge, but this is certainly a lot easier.

A few options to try.. Thanks
 

boriz

Senior Member
As a place to start, I like the translucent plastic idea coupled with three LDRs, each with an optical grade filter (Red, Green, Blue). Should be easily Picaxeable. Calibration would be a case of watching the sky and taking notes of the readings for different conditions.
 

boriz

Senior Member
Or just two LDRs. One 'plain', the other with a blue filter.

The plain one measures absolute brightness and the difference between that and the blue reading gives you the 'blueness' level.
 
Top