I’m no expert but as I understand it, light field cameras record both direction and intensity information about incoming light. As such, one of the things that you can do is focus after the photo has been taken! Through a microscope I believe that one can extract images that focus at different depths. Layering these images back together allows for the creation of images that are in focus at all depths.
I taught a class with kids where they collected benthic macroinvertebrates from river beds to look for sensitive and indicator species. We did this in three different places within the watershed, farm, forest, and suburban. It was a great success save for one major concern that the kids had. They didn’t want to put the invertebrates in alcohol. Standard pictures would not have had the depth of focus required to identify them correctly. We needed an expert as it was.
I’m wondering if one could put a light field camera (Or a regular camera modified to take light field pictures) on the openflexure microscope. Light field video or maybe a burst of images would allow us to focus after the fact and layer multiple images to get a larger depth of field. Under a microscope, they are much thicker than the typical depth of focus (not sure that I’m using the correct terms here. There is a “plane” where an image is in focus. That “plane” has limited thickness. When the sample is thicker than that region only part of the object is in focus.) Does this seem possible?
It would also make for some amazing pictures.
@Albercook that sounds like a fantastic project. I don’t know much about light field cameras, but one way to get the focusing after taking images is to just take multiple images at different positions in focus. The Openflexure software has a built-in stack-and-scan module that will automatically take a set of images at different positions and different focus.
It may struggle with a very thick sample, as to take a picture at focus and some pictures above and below there needs to be a reasonably well-defined position that is in focus. If you don’t need the x-y scan, then the class could easily use the Blockly extension to make a script to take a set of pictures at different focus positions after the user has found the mid-position (or more easily the lowest or highest position).
An image processing software like Image-J might be able to combine the stack of images into a 3D object, or help to visualise them in some other way, again that is not something that I have used.
How big are the organisms that you are looking at, and what magnification did you use on the microscope? Are they bright-field or dark-field images?
What an amazing idea. I think the OFM is anything you want it to be. Like William said you can do Z stacking of images captured at different focal points. Then use any Z-stacking software to blend them together. I use ImageJ from NIH or adobe photoshop with fantastic results. However, this requires a steady object and those organism will not cooperate. Another consideration is the processor power. I don’t think that the Raspberry pi has the graphic processing capabilities to be up for the task. Probably, is a more practical approach to convince the kids to use alcohol and z-stacking than trying to modify the scope at this point. Regardless, I think is a brilliant idea.
The little buggers are moving around too much to take multiple images and stack them later.
The alcohol is a non-starter for a large number of kids.
1 Like
A light field camera would be super cool, though probably also quite expensive and complicated. It may also not get the resolution you’d hope, because in a microscope you’re typically limited in resolution by the aperture of the objective.
I was also going to suggest focal stacking, though if you have moving samples clearly that complicated things! One possible option is looking at some of the work from Gemma Cairns, in Brian Patton’s group at Strathclyde University. They used multiple (4 or 5) cameras to acquire images simultaneously in different focus positions. It’s still more complicated and expensive than three basic microscope, but probably easier and better quality than a light field camera - and I believe they’ve already published open designs.
1 Like