Its either in street magic or the RC, but spirits do possess regular senses to use their Perception skill with. (someone in our party plays a free spirit, its come up a few times)
An astral being being able to use a Perception skill while Dual Natured without penalty is much the same as a a Mage using Assensing skill while astrally perceiving or projecting.
If you believe the Mage grows "Spirit Eyeballs" of the astrally perceiving variety while projecting or perceiving, then of course spirits can also grow metahuman eyeballs to perceive mundane things.
The ability to mechanically roll the dice for "Perception" does not automatically confer the physical organs of the perceiver. It just means they can use the skill.
"As with other dual-natured entities, a spirit’s ethereal senses are able
to sense both worlds without incurring modifiers for acting on
both planes at the same time."
"Spirits may use both Perception and Astral Perception skills
as normal."
But it doesn't explain why, it just simply says they can use the skill.
In fact, earlier in the general description of spirits.
"A free spirit, in its materialized form, is made of energy in
the form of a set of coherent forces. These forces effect the space
around it, giving it the illusion of shape, mass, and appearance."
and
"A free spirit has the “look and feel” of a physical
being, based on the form it took before it became free, but it is
not made of the materials that appear to make it up; “skin” will
not tan or sunburn, “iron” will not be attracted to magnets."
So while a Spirit can clearly operate while Manifest and use Perception skill, it's not because it has "regular" senses, but because it can reasonably interpret the mundane world.
As an additional item, if a spirit could use "regular" senses, why don't they work on Electronic Projections on screens? Wouldn't this clearly work if they did have a regular sense? It's just a picture isn't it? If they had regular eyeballs when Manifest there should be no reason it wouldn't work?
"Note, however, that spirits are unable
to see or interpret simsense, electronic projections on screens,
or AR displays."