Satellite maps are the most detailed maps we can have of our planet, and for most of time our maps have been partially incomplete or inaccurate, it’s only recently we’ve been able to have pretty much perfect maps.
When the first satellite maps were being made, were there any places that looked different from what we had mapped out, or any small islands and locations that we missed when sailing across the world?
Prior to the advent of satellite or aerial imagery, or global positioning, we did have methods of accurately mapping the Earth and the seas in various ways and to pretty high levels of precision. Especially with the advent of radio these methods could be even more precise because they allowed for the synchronization of clocks at long distances. So for the most part, any place that was known could be precisely mapped accurately by anyone willing to put in the effort. Though it sometimes took a great deal of effort. The surveying of the Indian subcontinent (The Great Trigonometrical Survey), for example, required hundreds of workers and decades of effort.
The biggest surprises that came when we first started viewing the Earth from space were ephemeral features. The large scale structure of clouds and weather patterns. Prior to the space age there wasn't a very good intuition of these things. Now, of course, we understand these patterns exist on a global scale, vast sweeps, sworls, and atmospheric rivers on the scale of oceans and continents. Similarly, we came to appreciate the seasonal variations in the land, the advance and retreat of snow and ice, or of green vs. brown.
Nevertheless, there were some less ephemeral surprises that came about from the first systematic satellite Earth observation surveys. In 1972 NASA launched ERTS-A (renamed Landsat 1 mid-mission) which was a modified Nimbus weather satellite design optimized for Earth observation. It was able to collect multi-spectral imagery of the Earth's surface at 80 meter resolution, that's nothing compared to what we can do routinely today but back then was pretty revolutionary. An 80m x 80m square is a mere 1.6 acres so you could collect imagery of the entire Earth down to a level of individual plots and city blocks. And it did produce a few surprises, specifically a small unknown island off the coast of Labrador in Canada, hence named "Landsat Island". The island is actually smaller than a single pixel for Landsat 1, just 25 x 45 meters across. It was officially charted as part of a 1976 Canadian survey. There were other minor revelations that came from the first Landsat survey including a few new lakes, other tiny islands, and a reef in the Indian Ocean.
Later satellite observations have led to the discovery of hundreds of tiny barrier islands (basically sand dunes that poke out above sea level), but as far as I'm aware no new substantial landmasses have been discovered.