It’s helpful, but if the task is to draw the imagined bicycle, then you would presumably add details as far as possible. But for most people the result is very far from what they would draw if a bicycle was in front of them. So what they imagine is not a bicycle. So most people can’t imagine a bicycle.
I can visualize things things well, but I'm aware that whatever representation I create isn't that accurate.
Like, I can't remember the details of every scene Catcher in the Rye, I can't recreate all of the notes of Freebird despite having listened to it a number of times, I and don't know the exact structure of a bicycle despite having seen and used them.
But I can still vividly "see" a bicycle despite this. It's more like my mind creates what it believes is a convincing bicycle based upon its most prominent characteristics. I can "see" it right now, gliding through a park (without a rider on top for some reason).
Perhaps if I were a bicycle mechanic, I'd be able to create a more accurate visual representation just because I'd be more familiar with the details.
I can only speak for myself but the more I gain knowledge/memory of an object, the more I can fill it in. If I know "bicycle" very well I can visualize all the details - maybe not in full detail all at once but at least by thinking about each part in turn. If my memory of one is not so good, the details will remain not filled in, like a blind spot where my mind fills in the gaps and I don't even notice the missing data until I try to "see" those areas specifically.
The imagined bicycle may be a specific one I recall, or a mixture of various memories of bicycles I've seen.