So, I can give a bit of a spread on what can be done there.
The posts appear to be botted, but they're coming in at rates indicative of a human poster making them. I can confirm this by looking at that big spam attack they did a couple weeks ago, it was entirely done on only two different IPs. Usually with an actual bot setup you'd use a different IP for each post, but as is it was really easy to clean up with a simple IP delete.
The solution to full bots is to implement captchas, which I can do very easily since it's built in the the LynxChan engine. But it would be annoying for regular users, and honestly the spam is pretty easy to clean up once a day or so (there's only ever a couple random posts here and there).
A filter wouldn't be effective because there's not much to filter when they're changing IPs and can just change their copypasta/image.
If you guys want to put on a captcha, go ahead and weigh in, but for now I'll just keep cleaning it up as it comes in.