@Tmaster I am not doubting the principle of increasing effective radiated power by way of relocating the antenna, or improving the radiation lobe in a set direction. The additional gain is important with weak signals at both send and receive ends, but at 50m it should not be an issue at these frequencies unless there are major attenuations to the signal. I have to guess you already have a dipole at the receiver ?
I use the Easy Newbie boards for most of my projects. It is designed for the nRF24L01+, and it has space for the needed 3.3v regulator if you are running it at 5V. I don't know if he has a way of using it with the RFM69 radios, but you could check. Plus it has a small proto-board area for wiring your sensors. Kudos to @sundberg84 for the design.
I'd say you present 2 switches, one triggers when someone enters the room, the other when someone exits, but without something else like a pir sensor that actually checks if someone is in the room or not, it will eventually leads to errors in the counter. All the counting will have to be made in the controller and maybe reset people count in a room if no movement is detected for like 30 minutes. How to do it in domoticz, I don't know 😁
I don't see many issues adding code to make it work with mysensors, but I'd use a Mega 2560 since the uno is really filled up and you don't have pins for radio; also you may be able to integrate the color sensor into the mega without using a second board. Keep in mind that you can that you can use reed or micro switches instead of color sensor if you are willing to add few cables running next to the blinds ☺️
Or add them if not present. Bus communication is nice but you need to know what you are doing, working with electric signals isn't like powering a light bulb. Once I had to explain to an electrician that rj45 plugs needed to have a specific cable sequence and he could just make one plug identical to the other.