Here's a correct answer: A craft couldn't.
We haven't mapped the whole of space yet so...
a craft with no other means of identifying its position
... can't know if it's near known or unknown pulsars. If you really have no idea of where you are or what your attitude is then looking around you isn't actually that useful. For example, you wake up in a forest, and from the flora and fauna you surmise that you must be in the Amazon rain forest. But you could easily be on another planet, in another galaxy, far far away.
Now if you have knowledge of your approximate position, say somewhere in the milky way, then you can start using pulsars and other things you see in space to calculate your position. You would need three known pulsars to triangulate your position (with 2 pulsars you can only know you're in 1 of 2 possible locations). If you know enough about the pulsars you could do some interesting red shift calculations to find out how far away from them you are. When it comes to the actual identification process, the duration of observation required is dependant on how many pulsars it might be, based firstly on your expected position and then on initial observations.
For example if you see a pulsar and you think you're near 1 of 10 pulsars then you can assume the pulsar you're seeing is one of the 10. Then you observe the pulsing of the pulsar and see at what rate it pulses. The longer you observer the pulsar the more accurate you calculation of it's spin can be, and hence the easier it is to rule out the other 9 pulsars.
The reason for a simple answer not being given to this question (by me at least), is that it is highly dependant on which pulsars you are looking at. For example take the following collection of pulsars and pulse frequencies:
It's clear that if you are trying to identify pulsar F it is going to be an easy task. If you see more than one pulse per second you can be confident that the other pulsars are ruled out. Now consider the following collection:
To identify pulsar F in this collection you Need to be able to observe the pulsar for around 13 seconds to rule out pulsar B as a potential candidate. This assumes that you can only count the number of pulses in a given number of seconds. This is unlikely, but there will be a limit to how accurately you can time pulses, dependent on the design of your hardware.
To surmise, the duration required to identify pulsar F in collection 2 is 13 times that of collection 1; this is why it's dependent on the number and characteristics of pulsars in range.
The following is based on values for 1,861 pulsars.
median: 1.88 Hz
highest: 716.35556 Hz (PSR J1748-2446ad)
lowest: 0.084897259 Hz (PSR J1841-0456)
This looks to me like a large number in the 1 - 4Hz range say around 700. So if the values are evenly distributed in this range there is approximately 0.004hz between each pulsar's equatorial frequency. The longest frequency involved in this large subset is 1Hz, which suggests that if you can measure to the nearest 0.004Hz then you can identify the majority of pulsars in 1 second or less. If your measuring accuracy is lower this will increase the time require.
Reference: The data is a composite from 52 published works. A list can be found at the bottom of the page HERE, under external sources.