Here is Buson's most metrically irregular (奇なる) haiku, according to Shiki:

をちこちをちこちと打つ砧かな
ochi-kochi/ ochi-kochi to/ utsu kinuta kana
"Here and there/ There and here/ Beating the fulling-blocks" (trans R. H. Blyth)

This one often turns up in discussions of onomatopoeia in haiku, probably because early Western haikologist Blyth used it as his first example of "[t]he direct representation of the sounds of the outside world by the sound of the voice". But I think that what Buson is doing is much more clever than simple bang-crash onomatopoeia.

You see, the thing about ochi-kochi is that although it looks a bit mimetic, etymologically it isn't; it derives from two OJ morphemes woti and koti which meant simply "far place" and "near place". And Buson uses it in this non-onomatopoeic way elsewhere, e.g. to describe waterfalls that are near and far. So the key to its use in this poem, the thing that makes it interesting, is the repetition.

The beating of the fulling-blocks is a notoriously rhythmic sound. There's a whole genre of shamisen/koto music called "fulling-block pieces", kinuta-mono, and their special rhythmic patterns are what set them apart. Skillfully played, they can induce an almost trance-like state. And that's what we see in this haiku, too: although it's broken up 4-7-5 by convention, there's no way to see the first "line" break on your first read-through. You just have to keep going: ochi-kochi ochi-kochi... until the to utsu snaps you out of it. The only overt sign of structure is the final five-mora closer (ending with kana, natch).

In other words, Buson uses the repetition of ochi-kochi to create a regular rhythm that is completely unlike what we expect to see in a haiku. The reference is not to the timbre or tone of a beaten fulling block, but to the rhythm, the strucure, of the beating itself. This might count as onomatopoeia in a broad sense, but it is certainly nothing like the poku-poku of Blyth's second example.