There seem to be a couple of different dimensions of variation in how human memory and imagination work. One is what is discussed here: the ability to "vividly" reconstruct remembered events "from a first-person perspective". Apparently, some people can do it, and others cannot. Presumably, there's a range of abilities in between.
At one end of the spectrum are people with hyperthymesia, who appear to be able to instantly reconstruct every tiny detail, including visual details from a first-person perspective, of everything they've ever experienced. At the other end are people like those discussed in this article.
Another dimension is how vividly and comprehensively people can construct, manipulate, and experience sensory models in their imaginations. Again, the spectrum seems to run from aphantasia at one end (that's the inability to construct mental imagery at all) to someone like Stephen Wiltshire, who appears to have an eidetic visual memory that enables him to draw large, detailed landscapes with roughly photographic accuracy after briefly seeing a scene a single time.
People with aphantasia generally seem to assume that when others describe mental imagery, they are being metaphorical; that they don't mean that they actually experience mental imagery in a literal sense. Contrariwise, I do actually experience mental imagery in a literal sense. In my late teens I became fascinated with a series of exercises devised by Aleister Crowley, in which you practice constructing increasingly elaborate assemblies of colorful 3D shapes moving in different ways. The goal was to see how many shapes you could manage at once and keep all of the shapes, colors, textures, and movement patterns stable for as long as possible.
I assume an aphantastic person would think Crowley was joking or bullshitting, until they learned that some people can indeed construct and manipulate detailed mental images.
Further complicating the subject, we also know from experiments in psychology of perception that the brain tends to edit memories, altering their content over time. I'd be interested to know how this tendency interacts with the variables mentioned above.
I'm somewhere in the middle on both axes discussed above. I experience memories in the first person, and I remember and dream in color. When navigating, I rely on a 3D model of where I am and where I'm going, and if asked to go somewhere I haven't been before, I want to see a map that I can use to refresh and supplement that mental model.
By contrast, my mother navigates procedurally. She wants a sequence of operations to perform: "turn right here; go a certain distance looking for such and such a building,..." and so on. I _can_ navigate that way, but I find it much more difficult and error-prone than just constructing the 3D model. From what my mother says, I'm not sure she can construct a model like that at all.
Signs of the difference show up in, for example, the kind of information each of us wants when being directed to a new place. She wants detailed step-by-step instructions; I want coordinates in space and some identifying information (like the street address) to confirm when I've arrived at the right place. She will follow directions to the letter. I'll just pick a route that points in the right direction and improvise until I'm close enough to look for the confirming data.