opencv - How can I separate color pixels that are not common for two images? -

here's do:

i have image of rice leaf. have image of rice leaf got brown color spots on leaf. want separate color pixels not common 2 images using opencv.(color of spots can vary)

i tried using histogram intersection. managed find number of pixels common between 2 images.

is there way using opencv? please me kind enough me.

if 2 images match perfectly

• if match use rhinodevel approach:
• so loop through pixels of first image
• and compare each pixel corresponding pixel second image
• if difference higher treshold
• you found not matching pixel , need do
• like add pixel output map or recolor (brown) pixel color first image or whatever

if 2 images not match

• so got reference leaf image , processed image can have position/rotation skew
• create list of colors per each image
• sort them ascending color
• cross compare booth lists
• if color in list2 not in list1
• then recolor/copy pixels contains such color in/from image2
• this approach slower o(xs*ys*n)
• xs,ys image 2 resolution , n number of non common colors

[notes]

• rgb fine may got better result on hsv color space
• in hsv can compare 3 parameters or few of them ignoring v value...