I recently read in a book that Canopus is sufficiently separated from the ecliptic to be used for stellar navigation and that spacecraft used to have "Canopus finders" on them for that particular purpose. Is it still being used this way?

Well - in a way. Most spacecraft still use a 'star sensor' and it is essentially the logical derivation of the Canopus sensor of the 60's in so far as it will identify whatever star field it sees and derive a pointing from that.

Doug

Today star scanners are used. The Voyager spacecraft tracked a single bright star approximately orthogonal to the sun direction for 3-axis knowledge. This makes the main body of the spacecraft need to be fixed in space -- not useful for probes which don't have scan platforms.

Basically, a star scanner has a map of thousands of stars and can recognize which stars it's looking at and hence reconstruct the accurate spacecraft attitude. The scanner is a wide-field camera that is deliberately out of focus a bit so stars are circular smudges on the detector -- this helps locate the star image centroids better than a single pixel would. They can provide very accurate attitude coordinates, but have a slight drawback when the real attitude slowly changes (for example during slews in target motion compensation) so one of the bright stars exits the field of view (or a new one enters) and the new calculated attitude (taking this new visible star collection) can suddenly "jump" to a new one, if only by a bit. This can then cause the spacecraft to "correct" for the perceived disturbance -- a jitter effect only noticeable in very precise tracking needed for example with narrow-FOV cameras and long exposures.