You would need to do much more! You need a shunt for ESD-events, so that the barrier in the DC/DC-converter isn't destroyed with large ESD-events. And these can be heavy large in a car!

Another disadvantage of DC/DC-converteres is, that you will notice the develop of lots of high frequency common mode noise at the secondary side! This is due to internal stray capacitance which transmits the fast edges of internal switching noise right to the secondary output! I have seen such an application failing the CE radiation test!! So, you must provide for the common mode noise a short path to the source of noise, means to the primary side, which can be done by connecting a cap from primary GND to secondary GND.
But this makes your ground routing complex again resulting in resonances at HF. The additionally needed caps destroy the isolation you plan to achieve with the DC/DC-converter for HF and again HF ground loops develop. And this even if you don't connect the moduls to each other, just because there's huge stray capacitance between your individual grounds and the chassis, adding complex impedances and making your filtering efforts rather unpredictable.
In an earlier post you mentioned that the chassis plane will inject noise into your circuits. Yes, but that's right what you get, if you let the chassis float!!!! Don't forget the stray capacitance! Everything is in close contact to the chassis of your car, just because the car is made of metall. Only if you connect the chassis to GND at as many places as possible you will prevent the chassis from injecting noise! You would need a car totally made of plastics, to achieve what you plan to achieve by your methode!

Mahmood contact an experienced EMC engineer and tell him what you plan to and I don't think, that he will find it a good idea to keep the chassis isolated and to use many DC/DC-converters.

Kai