From a historical point of view, essentially everyone in a regional area would indeed be aware of troops on the 'strategic' level. You may not know if the army will arrive at this hill or that river within a few miles (the tactical level), but you would know the general area.
This is due to a few factors. Firstly, the feudal system of the time was considerably more complex than the scope of the game implies; when a daimyo would issue an order to raise troops, he would have numerous vassals, who have their own vassals, who had their own to about 5 levels. That's a lot of people to be in on something and think it will be kept secret.
Secondly, they typically only waged war during specific parts of the year due to weather considerations. Marching in winter was a complete pita, and basically impossible in the northern part of the island. You couldn't depend on your muskets if the battle happened during the rainy season. If you levied up all your peasants during planting season, good luck eating that winter. If you were going to be invaded, it was going to occur within a fairly specific set of times.
Even on knowing the exact numbers of the troops, I bet they were fairly accurate on that during the period as well. All of the provinces around had already been rated in terms of income and manpower for many years. This process was called kokudaka. This is how the Shogunate functioned when it was strong; it gave large, prosperous demenses to more influential members of the court, etc. As a result, the relative value of different provinces, prefectures etc were often listed in terms of their value of koku and this knowledge was readily available. According to Wikipedia, the koku of Satsuma in Kyushu had a yearly income of 770,000 koku. A koku being enough rice to feed one person for one year.
Keep in mind also that during this period they attempted to consolidate the warrior class. Ikko-Ikki aside, things like sword hunts to confiscate weapons from the peasantry and making the samurai a hereditarily based caste system were more the norm. This makes the number of 'warriors' available even lower and more easily catalogued.
So you have a case where people often knew exactly how much manpower was available to their enemies, when their enemies would be able to attack them, and would have had knowledge of their mobilization quite early due to the distributed web based feudal system. Not to mention the fact that there were a lot of ronin, sohei, ninja, and other adventuring rumor monger types constantly roving around to help spread information even more.
Fog of war, at least the way it works in most strategy games, isn't justified. Nor do I think it's needed; what would it add? Isn't the uncertainty of the roll system in battle and the matchups between commanders enough randomness? I've lost fights where I thought I brought more than enough troops, and won ones I didn't expect to. Not sure how fog of war would really be an enhancement here.