Here is what I can share on Raspberry Pi and microscope objectives:
---I have worked on three different DIY microscope set ups:
-Public lab community microscope https://publiclab.org/notes/bronwen/05-07-2018/community-microscope-assembly-instructions
(SPECIAL THANKS to @cfastie for help with the 3D printed camera/objective lens holder!!!!! https://publiclab.org/notes/partsandcrafts/02-15-2018/2-attaching-your-raspberry-pi-camera-to-a-microscope-objective-lens
-Modular lens mount system
-substituting a raspberry pi camera in an old Celestron microscope.
This video shows a typical rebuild but I use a camera holder that was 3d printed : https://www.youtube.com/watch?v=M6TvCOPN5pI
---All of the systems above work equally well with Amscope or Celestron (standard) microscope objectives/ Raspberry PI cameras but there are some trade-offs:
1) Using objectives in non-standard configurations can reduce performance. An example is the use of a 30mm vs 160mm (standard) optical tubelength (see below).
2) Since the DIY microscope systems typically requires removing the factory installed Raspberry Pi lens, it is important to make sure dust does not reach the detector. If dust lands on the detector the choices are living with picture defects or risk damaging the detector while cleaning.
3)While objective lenses that I have used were built for visible light microscopy, the lenses generally work in the near infrared (with refocus) but not very well below 400nm. You can see some near infrared imagery here: https://publiclab.org/notes/MaggPi/06-21-2018/infrared-insect-microscopy Note I have tried 380nm diodes with my objectives but no light reaches the detector. Since I know the Raspberry PI camera has a 380nm UV response ( see: https://publiclab.org/notes/MaggPi/06-17-2018/how-can-you-use-computer-vision-to-reduce-spectral-overlap ) the glass used in the objective lens seems to be opaque to UV.
Examples below show the difference in magnification for the same camera (Raspberry Pi NoIR) and microscope objective (Celestron 20x) but with different tubelengths (30 vs 160mm).
The image is of human blood cells which is a good reference target since the cells have about an 8 micron diameter and a 1 micron center diameter.
The first slide shows human blood cells with a 30mm tubelength, 20x objective at two resolutions.
The second slide show that the 160mm tube length (with the same objective) provides increased magnification(left picture). The right picture shows that magnification with 10x objective/160mm tubelength is greater than a 20xobjective/30mm tubelength. The difference in magnification between 30 and 160 mm tubelengths seems to be a least 5 times and may indicate that the magnification is directly proportional.
So, I know I haven’t really answered the question about a low cost fluorescence microscope but I suspect any DIY design will have to compensate for using cheap optics. One possible compensation method may be computer vision software which can be designed to correct for image distortions.
Microscopy U offers a great explanation of specialized microscope objectives: https://www.microscopyu.com/microscopy-basics/specialized-microscope-objectives
@warren, @icarito, @amirberAgain