Best practices in standardizing scientific images
Here I have started collecting some very general advice on scientific imaging with the goal to create standardized images for phenotypic analysis. I am updating this loose and definitively non-exhaustive list as I go.
1 - Include a reference card - (almost) always!
A reference card in images can serve multiple purposes: on the one hand, a reference card is required to map size from pixels to real world units (e.g., from pixels to millimeters) and to calibrate color and brightness so that they become comparable across images and photographed scenes. On the other hand, including a reference card in every images will ensure that the images are comparable to each other - e.g., in the case of varying brightness or distance from camera and photographed specimen. Only if you have completely invariable imaging systems, like scanners, plate readers, and some microscopes, you are probably fine to not always include a reference card.
See my post on how to create color and size reference cards.
2 - Dealing with light reflections (this is for the aquatic ecologists)
When working with aquatic organisms you may want to photograph them in water to not harm them, if they are alive, and to get a better representation of their phenotype. However, this can create new problems, such as reflections. When working with an imaging system that photographs the specimen from the top. To avoid this, you can add diffusor-screens that scatter the light from the source into smaller rays. Another approach is to simply place your light sources in a fashion that directs the outgoing light not directly into the lens. Just like playing billiard: angle of incidence is equal to angle of reflection.