Flora,

you are not posting the full code, but only an image of only part of the code. One problem seems to be that you are trying to detect the spots using ColorsNear[Black] - but having a black background.

Here comes another approach - basically as an one-liner:

img0 = Import["https://community.wolfram.com/c/message_boards/get_message_attachment?messageId=3396992&attachment=Screen+Shot+2025-02-19+at+6.40.56+PM.png"]
HighlightImage[img0, ComponentMeasurements[Binarize[ColorDetect[img0, ColorsNear[RGBColor[.2, .25, .25]]], .6], Centroid", 10 < #Count < 2000 &]]

to prevernt showing points outside the image,what should i do?

A simple method could be to use a mask defining the leaf shape - and than working with the resulting image, i.e. like so:

img0 = Import["https://community.wolfram.com/c/message_boards/get_message_attachment?messageId=3396992&attachment=Screen+Shot+2025-02-19+at+6.40.56+PM.png"]
leafMask = FillingTransform@MorphologicalBinarize[img0, .3]
img1 = leafMask img0

As always: You have to experiment with WL-functions and their parameters - but its fun!

Flora,

the function Binarize[] has a parameter! In this case the value ".1" is much too small - just play around using other values, or do not use it at all.