Feature detection on a small, noisy image with OpenCV -

Feature detection on a small, noisy image with OpenCV -

i have image both pretty noisy, little (the relevant portion 381 × 314) , features subtle.

the source image , cropped relevant area here well: http://imgur.com/a/o8zc2

the task count number of white-ish dots within relevant area using python happy isolating lighter dots , lines within area , removing background construction (in case cell).

with opencv i've tried histogram equalization (destroys details), finding contours (didn't work), using color ranges (too close in color?)

any suggestions or guidance on other things try? don't believe can higher res image task possible rather hard source?

(this not python answer, since never used python/opencv binding. images below created using mathematica. used basic image processing functions, should able implement in python on own.)

a general "trick" in image processing think removing thing you're looking for, instead of looking it. because often, removing much easier finding it. instance apply morphological opening, median filter or gaussian filter it:

these filters remove details smaller filter size, , leave coarser structures more or less untouched. can take difference original image , local maxima:

(you'll have play around different "detail removal filters" , filter sizes. there's no way tell 1 works best 1 image.)

opencv image-processing image-recognition edge-detection noise-reduction