If I were doing this, here are the motion settings that I would change from default (looks like you have your zone set up pretty well so I'll skip that):
- Units -> Pixels
Alarm Check Method -> Filtered Pixels or Alarmed Pixels (experiment to find what works best)
Filter Width/Height -> If using Filtered Pixel method, must be smaller (in pixels) than the smallest object you want to see
Min Alarm Area -> Something very small, say 10 pixels for starters
Max Alarm Area -> leave empty for now, but you can use this to filter out larger objects you may not be interested in (e.g. when the moon comes out from behind a cloud)
Min/Max Filtered Area -> Same as Min/Max Alarm Area
This list is by no means exhaustive and is simply an educated guess on my behalf. Consider it your starting point. You need to experiment to find the best balance between capturing the objects you want vs. capturing noise or other unwanted objects. For example, Alarmed Pixels may be better at picking up the small objects you are interested in, but that will also make it more sensitive to noise.
Also, if you have not done so already, I recommend you spend 15 minutes and take a look at the documentation. It spells out, fairly precisely, what all the motion detection settings do. I know I still have to reread each description several times just to refresh my understanding of what the author means.
Lastly, if you continue to struggle, please make any future responses more quantitative. Provide us with all the motion settings you are using and even the make and model of camera might help.