I see several constraints to the size related to the space environment, spacecraft design, as well to scientific instrumentation. Although micro-probes are being considered, in space, EVERYTHING is a trade-off. I'll start first with the restrictions, but I also add some details after that about what is currently enabling those kinds of probes.
Radiation outside the Earth's magnetosphere is relatively high. A considerable amount of cosmic rays that are usually deflected by the magnetic field will hit your spacecraft. This causes several headaches to spacecraft designers, but one of the most prominent is that the reliability and lifetime of electronics drop considerably. Something like a raspberry pi or any regular consumer electronics has no hope to survive in space for any useful amount of time. Effects of radiation on electronics
You can improve your mission chances of success by adding considerable shielding which will definitely add weight, but most importantly, you MUST use radiation-hardened electronics, which are neither cheap nor anywhere as diverse as regular commercial products. (for example something like this: Rad Hardened FPGA). And this does not only applies to semiconductors, but also to PCBs, batteries, data busses, and communication systems, so you either buy the spacecraft components available on the market, or you build custom ones, which are even more expensive.
To give you an idea, the cheapest onboard computers for CubeSats cost around $5000 and will survive maybe 3 years on Low Earth Orbit which is highly protected from radiation compared to interplanetary space or other planets. Something like the Juice probe that is soon to be launched to Jupiter and will travel for 6 years before arriving, needs to survive for decades in space and requires very specific and even custom components to achieve its mission.
Added to all this, because of the manufacturing processes needed for radiation hardening, those electronics are also bulkier and less capable than their regular counterparts.
Then there is the temperature as well, consumer electronics are made to operate at a very mild range of temperatures (-20 to 40 degrees Celcius). Space electronics must be capable to operate in considerably higher ranges. (-40 to 85 degrees Celcius with proper thermal management in the Spacecraft), because in space when the sun hits you is REALLY hot, and when it doesn't is REALLY cold.
Then you have the payload part. You mention a camera. Well, any camera needs to be radiation hardened as well so you can't simply put any commercial device, but you must also ask why you want to send a camera.
You can take pictures for sure, but the resolution is highly reduced, if you want better resolution for example to study geological processes on another planet, you're going to need better bigger equipment. But most of the time, the scientific payloads are not cameras in the commercial sense.
Yes, you can have pictures that are useful, but most payloads are designed with a very specific type of research in mind, we're talking about ground composition studies, searching for subsurface water, etc. Those kinds of payloads are not small at all. The Mars Reconnaissance Orbiter has like 4 different types of cameras for very different types of data collection, a spectrometer, subsurface radar, etc
(MRO FactSheet), and those are big, shoebox big.
As the payloads increase in size, also the power requirements, which will increase the battery size, solar panels... Sometimes you even need shielding between payloads so one does not interfere with the other.
The reasoning with most missions is that if you are already sending something, let's try to maximize the useful mass of your spacecraft.
Another limitation I can think off is communication. NASA uses the Deep Space Network, which consists on 3 sites around the globe with huge dishes, to communicate with anything that is far from the Earth. They even plan to use it for Artemis. But the probe must be able to reach them, there are limitations regarding the power of a transmission amplifier and the gain of the antenna, and their size, so if you go too small you risk not being able to communicate at all. Yes you can do deployable antennas, but that adds complexity and risks to the mission (design, testing) and makes it most expensive.
Now, one important point is that nanosatellites didn't really become mainstream when it comes to science until very recently. There was no availability of components or payloads to build such a device, but that has changed. Along with Artemis 1, there are 10 CubeSats that will be deployed around the Moon to perform scientific experiments Artemis Cubesats. Along with NASA's InSight Mars lander, there were 2 CubeSats that were designed to act as relay satellites for a technology demonstration mission (MarCO). ESA's Hera mission will also have 2 CubeSats that will collect additional data (Hera). So nanosatellites are starting to be considered for this type of mission.
But so far there are restrictions, you usually would need to stick to only one payload, if you want very specific data, probably a custom payload that will take years and a lot of money to design, test and manufacture. Sometimes you would need communications relay infrastructure in place, for example, the MRO could be used as a relay on Mars.
Then, due to the small size, propulsion is really an issue, is not only ISP, you need the right thrust as well. You cannot perform an orbital insertion with a very high Isp engine if it takes a month to actually achieve the required DeltaV (for example ions, which are very efficient but very very slow). The selection of thrust and Isp is also a trade-off, increase one and the other one will decrease, which limits the actual useful alternatives. Right now most of these missions have very low to no propulsion at all and rely on hitchin' a ride with other missions to get to their destinations and have very reduced maneuverability.
Things are changing of course. More and more research and development are going into micro and nanosats, and I expect these types of missions will increase considerably with time.
One last thing, you can probably send 200 probes like the Falcon 9 example you mentioned, but that only makes sense if you NEED the 200 probes, like in some very specific type of experiment or data collection mission, but if you can collect the same data with a single bigger spacecraft, then by far, is more efficient to send only one.
I hope this helps!