How to distinguish Probe GET from Legit GET requests

Is it possible to distinguish between unwanted GET requests with path-names but without website URL and normal/legitimate GET requests containing Full URL names?

Reason for this inquiry is that I see a behavior pattern in my HAProxy logs the starts with GET requests without a URL name, followed flooding of GET requests probing(guessing) all kinds of path names related to apps like WP / Wiki / PHPmyAdmin, Etc…

Many times, this behavior is repeated in a coordinated manner from numerous different IP addresses almost as if the first IP informs the others to join the scan(probing).

My goal is as follows:
REJECTGET /pw/Main_Page HTTP/1.1
ACCEPT “GET https://mysite.com/pw/Main_Page HTTP/1.1”

Is this something feasible with HAProxy ACLs?

P.S.
I am not worried about googlebot or other crawlers not being able to index my site.

Thank you.

Just check the host header e.g.
acl mysite { hdr(host) -i mysite.com }