I've read a certain book, well known to the researchers of history of the WW2 Eastern Front, which advocates the idea that USSR had planned a massive full-scale offensive against Nazi Germany set to just two weeks after the Germans actually struck their preemptive strike on June 22nd. Not discussing the assumption in general, one particular account bothers me ever since. The book's author discusses that all soviet units poised for invasion were supplied with large amount of high-quality military-grade topographical maps of the German-controlled territory. And that when Germans attacked, these maps quickly lost their relevance but topographical maps of domestic areas, vital to mount any defensive effort, were quite scarce in comparison.
So I would like to know how the maps are obtained whenever one's general staff plans to invade some rather unfriendly country. I would assume the maps have to be fresh, reflecting most, if not any, recent developments of foreign infrastructure such as paved/unpaved roads, railroads, tall structures, bridges, dams etc etc. But how? If the maps are done by some military agency, then they are classified. If they are done by independent contractors then I would assume some kind of secret service to be monitoring their clientele just for this reason. There are no any space technology yet. Use of spy planes would be logistically unfeasible and may hint the foreign intelligence to our plans. Use of actual spy network seems too be risky and cumbersome. And yet somehow this was done on a regular basis. How?
One of the side-projects related to my work (language documentation) is digitising old survey maps. I've basically taken every single survey map of the area I can find, and then painstakingly aligned it with modern GIS data — USGS Digital Elevation Models, Open Street Map where it's applicable (which is basically never unless I'm the one who's put it in, it turns out) — and in doing so try to account for what was done insufficiently well in the past. The area I'm most active happens to be the old China Burma India Theatre of WWII, along the Patkai mountain range and the Pangsau Pass.
Starting in the 1800s the British made great efforts to survey the area, and there are all sorts of great accounts of such survey missions published at the time and still available in the second-hand bookstores of Rangoon today.
In the late 1800s, the maps of this area were terrible. Huge areas where guestimations based on what could kinda be seen, or what locals had told the British. The words for "mountain" and "river" and "stream" change all over the place based on the language spoken by the guy they were apparently asking. You'll have two villages marked on an un-surveyed ridge that have similar names, which if you know the languages spoken in the area quickly become apparent as the same word just said by speakers of two different languages.
This improved greatly over the decades. The 1920 series of maps are way better than those from the 1890s expeditions. The 1927 maps correct a lot of mistakes in the 1920-1921 series. 1932, 1934 and so on saw other editions. In large part this was just because it took that long to conduct a survey, so between 1932 and 1934 a big difference might be visible in one area but no difference in another, but actually both started out in the same shape. It's not until the 1930s and 1940s that things really start getting well surveyed in this particular area.
To me at least, the really interesting thing for me though, is that around this time, Japan suddenly has good maps of the area too. That's because it was around that time that the Japanese were simply reprinting the British survey maps.
These maps simply existed out in the world, even though some editions of them were fairly restricted in who had access, they weren't that hard to get ahold of, and during WWII the Japanese did exactly that. Maps that are clearly not just pasting-over the English with Japanese but obvious reproductions were produced.
Since the publication of them and the surveys conducted to gather the data all took so long, so much of the information was already out there by the time the Japanese had taken Rangoon and Joseph Stilwell was doing his thing in India. Japan obviously hadn't been in the region long enough to do their own surveying, but getting a paper copy of the data wasn't as hard as cracking secure high-res satellite imagery databases or anything like that.
Also of note is that all the Japanese maps which are clear copies of British maps were published about 2 years later. My 1934 map from the Brits matches the 1936 Japanese map, and the like. I'm traveling so don't have access to my map collection to give exact examples, but I know it's about one to three years between the British version and the Japanese copy in each case where I'd checked.