This is a cheap and optimistic way of extracting numbers from text. It is quite easy to think of cases where this won't work, but there are tons of real-life cases where it works nicely. I allow for numbers at the end of sentences, like 1.23e3. I certainly don't handle expressions like 1.2+3.4-5.6. This was chosen for simplicity of the regex.

Update: Now also handles cases like "123ears" after reading Identifying and parsing numbers.

my @numbers= $text =~ /([-+\d.eE]*\d)/g;

Previous version was:

my @numbers= grep { s/\.$//; /\d/ } $text =~ /([-+\d.eE]+)/g;

Update: petral's excellent advice incorporated.