Analyze (count) non-linear line segments with similar colour to background?


I am new to image J and am trying to automatically count objects in a series of photos. The photos are of a known quadrat space with a number of animals inside. The issue is that the animals (non-linear line segments), are a similar colour to the background. I have attached an example. I cannot seem to figure out how to get ImageJ to deal with this kind of photo because there is very little difference between the objects and the background in terms of colour or intensity (when in grayscale).

Any help would be very appreciated!!

Have you tried infrared photography?

I assume the animals in your photo rest on algae. Most green plants reflect a lot of light in the infrared, appearing white on the photo, while animals usually absorb more light and will appear darker.

To make a test you just need an infrared torch, a cheap webcam and remove the ir-blocking filter. There are many (video)tutorials on how to make this. If you can choose, it is best to use 940nm infrared leds compared with 850nm.


I will give this a try. Thank you!

Basically you have two options: remove the camera ir-blocking filter and make the photo in the dark with IR light or add a band-pass filter and make the photo with sun light or a normal incandescence lamp.

Public Lab has many tutorials on infrared photography and hardware modding: