7
\$\begingroup\$

I'm looking for sensor options for proximity detection through a pane of glass. Something that will sit 6-24" on one side of a pane of glass, and be able to detect presence the same (roughly) distance on the other? I assume ultrasound won't work, but am not sure about infrared. If not, are there other options?

Edit per this link from the comments, most glass is opaque to IR, so no, it won't work. That leaves the second part of the question - are there ways to detect presence / motion on the other side of a pane of glass. Video-based motion detection won't work because I need to be able to restrict distance (i.e. something present/moving within a fairly tight distance of the glass). Use case is something along the lines of a home security motion sensor looking out the window.

\$\endgroup\$
13
  • \$\begingroup\$ There is some useful information in this thread if that helps. \$\endgroup\$
    – Roger Rowland
    Commented May 21, 2015 at 14:59
  • \$\begingroup\$ Yes, that's helpful - thanks. Tl;dr - IR won't work through glass. That leaves the other part of the question then - how to do proximity detection through glass. \$\endgroup\$
    – kolosy
    Commented May 21, 2015 at 15:01
  • \$\begingroup\$ IR can work through glass, it just depends on the glass, the type of IR, and the angle/relation between the two. Is this a situation where you could have a laser through the glass? Would only work if you knew where the object would be passing through. \$\endgroup\$
    – I. Wolfe
    Commented May 21, 2015 at 15:03
  • \$\begingroup\$ @I.Wolfe no, unfortunately I can't pass a laser through, and don't have control over the type of glass. This is something along the lines of a home security motion sensor looking out the window. \$\endgroup\$
    – kolosy
    Commented May 21, 2015 at 15:04
  • 5
    \$\begingroup\$ Don't confuse the different wavelengths of IR. 850 nm IR, the kind you get from an IR LED and detected by a photodiode, will pass through most glass unhindered, just like visible light. 12000 nm thermal IR, the kind emitted by a person and detected by a passive infrared bolometer, does not pass through glass. See Reply 13 of the thread you linked to. So how about a reflection sensor based on IR LEDs? \$\endgroup\$
    – tomnexus
    Commented May 21, 2015 at 15:47

5 Answers 5

Reset to default
3
\$\begingroup\$

You could use a microwave presence sensor e.g. this one which mentions glass specifically

In general, they will work through glass, without a big reduction in range.

They are often combined with PIR detectors to reduce false triggering (from sources of heat), but usually configurable so you can only use the microwave part.

\$\endgroup\$
2
\$\begingroup\$

IR goes through glass if it’s anywhere near the visible region. Soda-lime glass:

enter image description here

\$\endgroup\$
1
\$\begingroup\$

This kind of situation is usually dealt with using video processing. A camera pointing out through the "window" and comparing successive frames.

You can either do "advanced" object detection (there's a big open source project just for this kind of thing: http://opencv.org/) or take a more simplistic approach which I have used myself:

  1. Maintain a "buffer" of the past (for example) 10 frames
  2. Identify any pixels that have changed between frames
  3. Count the percentage of pixels that have changed
  4. Calculate the number of successive frames that have changed pixels over a certain threshold.

If enough has changed then you have motion.

You can also split the image up into "zones" so that you have only one area that is sensitive to motion, or make some areas more sensitive to motion than others.

\$\endgroup\$
1
  • \$\begingroup\$ How can you calculate distance to an object with video? \$\endgroup\$
    – Abdel Aleem
    Commented Jan 14 at 1:10
0
\$\begingroup\$

Induction would work, I believe. IR doesn't work because of the glass, yes, but close range coupled coils can still sense each other, and the voltage drop across the primary is inversely proportional to the distance of the secondary.

Coupling gives you a means of differentiating receivers, voltage drop gives you a means of measuring distance.

If it's for security though, it probably not going to be the case that some robber will carry a tuned receiver coil, so perhaps have an IR section hidden on the outside, fed power by induction and sending data back by induction too.

Food for thought.

\$\endgroup\$
2
  • \$\begingroup\$ I don't think it would. One, it doesn't fit the use case at all, two, what's the effective range of an inductive setup like you describe? a few inches at most? \$\endgroup\$
    – kolosy
    Commented May 21, 2015 at 17:44
  • \$\begingroup\$ Depends on the frequency, and for further distances; the distance of the near field. You aren't meant to use the inductive coils to 'detect'. You use just them to allow you to put the IR transceivers outside of the glass. \$\endgroup\$
    – ARMATAV
    Commented May 21, 2015 at 18:27
0
\$\begingroup\$

It is glass. You can use visible range, right? If your purpose is to detect someone's presence, have a camera which does it for you. Motion is a great project to achieve this.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.