Having done both, I think the kernel community and academia are surprisingly similar.
In both cases if you are only proposing an incremental idea/patch, it's fairly easy to get that looked at and considered.
However if you are planning something drastic, expect months to years of "peer review" (be it by entrenched professors or else just no-nonsense kernel devs).
The process can be long and arbitrary, and it turns off enough people that they don't bother trying. And yes, you have to fight politics, on both ends, and outsiders are treated with suspicion at first.