A "cheap low-res IR camera" isn't really a thing. Thermal imaging cameras cost a lot more than your average microwave, even at resolutions we would consider "su

A "cheap low-res IR camera" isn't really a thing. Thermal imaging cameras cost a lot more than your average microwave, even at resolutions we would consider "super low" by visual standards. The only way it would be worth it is if it worked flawlessly, which it probably wouldn't even with a pile of R&D. Microwaves intrinsically never heat food evenly, so in order to heat the whole system to your desired temperature, you must perforce heat parts of it to above your desired temperature, or switch the magnetron on and off an excessive number of times as the system asymptotes. Nor are the hotspots in a random place, but are fixed resonances; the best we've been able to do about that is a turntable, which turns a hot spot into a hot radius. So you're going to be heating the same spot multiple times. Microwaves are just unsuited for this, and a thermal camera won't help much. For your poaching salmon application, you probably want a souse vide bath. You can build a passable one for pretty cheap - it's just a pot of water on a hot plate and a thermocouple. It will heat the entire surface of the food to exactly the target temperature and hold it there; you can detect when the food is up to temperature by measuring how much energy you need to maintain the bath. To boil water in the microwave, may I suggest a sealed container with a plastic whistle, like a kettle?