The DC-DC converter is an excellent idea. But it’s best to make sure it can produce the correct output to the battery for a range of panel input voltage from both above and below the final battery voltage.
However, a simple single-inductor Buck switcher can only drop the voltage.
SEPIC or Zeta switchers each need two inductors but most importantly, they can cover the possible input range and still produce useful output. Often these still use a Buck switch controller.
The blocking diode is best on the lead from the panel, not on the lead to the battery. This is because the diode’s forward voltage will vary with current.
I’ve done several experiments with solar panels, most recently during lockdown. I used the OBDII socket inside the Mazda3 to charge the car battery because this is not switched off with the ignition.
If I put a panel up inside the windscreen (so it wasn’t nicked) my results suggested I needed one about the size of the windscreen in the winter to have a net gain in Ampere-hours per day over dark current!
The “2.5W” NOCO panel I have is effectively useless and not on sale
I dug it out during again during lockdown in a vain hope it might keep the Mazda3 battery alive. No.
I tried again with a “10W” panel. Also no joy.
This “100W” panel has almost enough ampere-hours per day in the depth of overcast winter, but it’s too big for inside the car, and without a suitable regulator (and blocking diode for the dark) it could boil the battery if well aimed on a really good day! I gave it 3* in my review.
If anyone wants it for further experiments they can have it, but collection only (PM me).