UDI R/C U816 / U818 Protocol

you can hook the probes to the MCU which likely has much wider pitch (I can see it in the bottom of your pic)
Alternatively, you can (carefully) scrape off the covering on the traces and solder to those. I don't like doing so, but sometimes it is the easiest way.

I don't know if btoschi ever got the UDI818 working. As he notes above there are 2 different models with different protocols. Is there a reason you chose this specific quad? If you are focusing on machine-vision, why spend time developing a protocol 1st? Go to the list of supported models on the left of your screen, pick one, then golook at the arduino code posted here: www.deviationtx.com/forum/6-general-disc...ones-with-smartphone

If you want help here, you need to ask specific questions. You mention a 'transceiver' but in RC, we have a transmitter (the thing with the sticks) and a receiver (the thing in the vehicle) regardless of whether there is bi-di communication. I don't know what issues you are having, so I have no idea what pointers you need. I can tell you that I am unlikely to try to debug someone else's code on Arduino, given that I primarily work on ARM, so if you are looking for a set of eyes, they probably won't be mine. But if you need help on methods, or spi parsing, then I can try to help.

it really depends on the vehicle. In this case btoschi already tried with limited success I believe to analyze the protocol, so this is probably not the best choice if protocol analysis is a secondary goal. My point was purely that if your focus is machine vision, why get side-tracked doing protocol analysis which could end up as its own rabbit hole.

That said, technically, yes hooking up an SPI sniffer to the Tx is the best 1st step. You need to determine the initialization sequence of the Tx, and the timing and composition of the packets. Generally these are reasonably easy to do. The initialization sequence can just be replayed. The packets are usually 16 bytes unencrypted with 1 or 2 bytes per channel. With an NRF based protocol, you need to worry about whether there is a bi-directional binding scheme, and whether any data from the Rx is used in the channel hopping sequence. Usually the channel-hopping algorithm is the most challenging to figure out. it may be a hard-coded set of channels (which may or may not be transmitted to the Rx ahead of time), or it may be computed based on the Tx and/or Rx ID. Often it is not too difficult to emulate a specific Rx/Tx pair (y replaying the Tx channel hopping scheme(, but then you can't fly multiple vehicles in the same airspace because you need to know the parameters that determine the hopping.

Again, given your goals as I understand them, I think you'd be better off with a different quad as the basis, but if you really want to go forward with this, then read this thread carefully, look at btoschi's existing work, scan the SPI bus and see how your data compares to his. And post here any questions since we will certainly want to incorporate any of your findings into an eventual Deviation protocol.