0
$\begingroup$

I'm currently tagging insects with reflective tags and filming them in night-mode on my camera which has an infrared light. I have some markings on the tags which I need to catch on film. Unfortunately, when I film about 2 metres back, the the contrast between the reflective tag and the marking is too great, resulting in the white reflection being the only thing visible. Does anyone have any ideas for how to fix this problem? For instance by using filters?

enter image description here

enter image description here

Improve this question
$\endgroup$
2
  • $\begingroup$ An illustration of the problem would be great. $\endgroup$
    – Emre
    Nov 2, 2014 at 6:52
  • $\begingroup$ Good idea, just added some :) $\endgroup$
    – Jack Simpson
    Nov 2, 2014 at 14:05

2 Answers 2

Reset to default
0
$\begingroup$

Reducing exposure time should be enough to work with here. Since you're video recording their movements (from my experience) there isn't much else you can do unless you can get a much higher resolution as well.

Just an idea for an alternative way of tracking, (and I believe I'm treading dangerous waters here because

1) this is an idea and I don't even know what you want to do with the video, so I chose tracking as a possible application

2) not sure if it complies with the rules of answering a question)

If you can take a picture of the first shot with a high res camera (change exposure time, shutter speed etc.) and then immediately record a video with your hardware after that, you should be able to see the tags on the insects clearer in the image you first took. If you want to track patters of a specific insect, perhaps extract the location of the corresponding tag in the image and the insect would be very close to the same location in the very beginning of the video (if not at the same spot) Then you should be able to understand where which insect is in the video and track them. The caveats here would be

i) Taking a low exposure high res image first and immediately taking a video after that

ii) if there's a cluster of insects all close to each other, you'd have to manually verify that the location in the video corresponding to the one you chose in the image has the right insect. Else you might end up tracking the wrong insect with the wrong tag.

iii) involves processing after capture of the entire video (or real time possible too)

Improve this answer
$\endgroup$
3
  • $\begingroup$ Thank so much for the advice, definitely agree about the exposure, I tried to block off some of the IR light and it helped the problem a fair bit this afternoon. I'm interested in tracking them so I can compare the behaviour of different groups, and I agree having super high resolution image taken every now and again would be great, except it would be a really expensive camera to do this and part of the idea of the project is to create a system that isn't too expensive to replicate. $\endgroup$
    – Jack Simpson
    Nov 3, 2014 at 10:02
  • $\begingroup$ @JackSimpson Just out of curiosity, are you using the same wavelength/frequency for each tag? Meaning, if say you have Tag1 for insect A which gets brightly illuminated for F1 frequency in IR, and Tag2 for Insect B which gets illuminated at F2 frequency, You might be able to track them by filtering out the corresponding frequencies during processing. Of course, while this might solve the exposure problem, you'll be faced with a new one of filtering out data in the video later. In case you're not familiar, check out Hyperspectral Imaging and see how they go about things. It's really fun stuff. $\endgroup$
    – Srihari S
    Nov 3, 2014 at 14:12
  • $\begingroup$ The tags all use the same reflective material, to be honest finding a reflective material that worked well with IR light took a while and I don't know where I'd find material that was specific to different IR wavelengths. $\endgroup$
    – Jack Simpson
    Nov 4, 2014 at 3:47
0
$\begingroup$

Use your camera's RAW mode, if it has one, and underexpose. If these concepts are not familiar to you, head over to https://photo.stackexchange.com/

Is the first picture a close-up of the second one, because the letter K on the tag is quite legible? If not, you may also want to use a higher resolution camera, and/or not shoot from such a distance.

If you think about it, you are asking us to robustly encode roughly four bits (on letter of the English alphabet), judging by the first picture. You could simply map the binary representation of the letter to five boxes; four in a square and one on the side, so you don't get any confusion due to symmetry. (If you laid them side by side you wouldn't know which side is which.) If you need to encode multiple letters, you could try using QR or a barcode.

Your first order of business has to be getting the exposure under control. I'd investigate the possibility of using aftermarket firmware if the manufacturer's doesn't give you enough control.

Improve this answer
$\endgroup$
1
  • $\begingroup$ Yep, the first photo is with the camera closer to the tag. I'm using the Sony AX100 - a 4K video camera, although unfortunately when you use it in nightshot mode with the IR light, it prevents you from manually changing the exposure levels. $\endgroup$
    – Jack Simpson
    Nov 3, 2014 at 6:01

Your Answer

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

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