51–60 of 67 posts
They've used a vision transformer to estimate building heights from monocular aerial photographs, so they're guesses at best. Calling this a map is a stretch.
I checked the building heights where I live vs the actual heights, sometimes they display the height as half of what the actual height is.
What is the crowd process to upgrade/correct this? I notice massive errors in small Midwest US towns (multiple buildings being combined/ommitted, etc)?
It's machine-learning generated "slop" honestly. Looking at where I live and where I grew up the building heights are quite badly estimated. - Some groups of houses around here that are more or less identically built but on sloped terrain are reported to have widly differing heights - My neighbour building is reported to be half the height of this building (they're more or less equally high at 5 stories) - A small of…
The height measure when you click on a building is not very accurate. Clicked on two next to each other (one single floor and one with two floors) and the single story house was marked taller by the site.
The positioning of the shapes in relation to each other are also wrong. And not in a subtle way.
Looked at the house my father built in 1985. A few hundred kilometers from my current location. The shape is wrong (as are the shapes of all neighboring buildings. As are the positions toward each other (distance between houses, rotation of shapes). The heights are also significantly wrong. The two story houses on the opposite side of the road are said to be slightly above 2 meters in height.
The height measure when you click on a building is not very accurate. Clicked on two next to each other (one single floor and one with two floors) and the single story house was marked taller by the site.
Earlier quoted context omitted.
It's machine-learning generated "slop" honestly. Looking at where I live and where I grew up the building heights are quite badly estimated. - Some groups of houses around here that are more or less identically built but on sloped terrain are reported to have widly differing heights - My neighbour building is reported to be half the height of this building (they're more or less equally high at 5 stories) - A small of…
Or, they subtracted a digital elevation model from a digital surface model, ran a point-in-polygon match against an existing building dataset, and labelled the difference as the height of the building. No ML needed.
What is the crowd process to upgrade/correct this? I notice massive errors in small Midwest US towns (multiple buildings being combined/ommitted, etc)?
It's machine-learning generated "slop" honestly. Looking at where I live and where I grew up the building heights are quite badly estimated. - Some groups of houses around here that are more or less identically built but on sloped terrain are reported to have widly differing heights - My neighbour building is reported to be half the height of this building (they're more or less equally high at 5 stories) - A small of…
They've used a vision transformer to estimate building heights from monocular aerial photographs, so they're guesses at best. Calling this a map is a stretch.
So here is finally the data that I needed for my idea: walk directions from A to B while never leaving the shadows, suggesting bars and pubs where to wait while the shadows catch up and let you cross “safely” where before was sunny. Useful for scorching weather places like south of Spain.