So this whole week I've been working with this computer program to help me count pumpkin pollen grains super fast. In a few seconds I can process the image of a pollen slide from our microscope camera and get a count. This is compared to 1 hour for 1 slide when counting by hand. The program creates a high contrast image, which makes the darker pollen stand out against the lighter background and then it turns the image into binary black and white to count the number of black blobs against a uniform white backdrop. Here is an example of a before and after image. Right now, the program over-estimates because it picks up debris and shadows, but I'm trying to figure out the best process for setting up the microscope slides so that we can reduce all of the mess and clumping.
I tried to do this with caddisfly larval case grains; does NOT work well for them because all the grains are different colors. I could never adjust the contrast in a way that some of the grains weren't diluted. Ugh.
ReplyDeleteWhat program are you using?
I am using LAS EZ for taking the microscope photos, and ImageJ for manipulation and counting. ImageJ is free, and super awesome if slides are perfect. Works on all operating systems too.
ReplyDeletehttp://rsbweb.nih.gov/ij/
Did you trying dying the casings?
Oh, I know ImageJ well. Love that it's free, too.
ReplyDeleteThe only problem with dying the casings is that I was also identifying the minerals, so color was important.