Segmentation of houses and roads on maps
50 masks + originals
LabelMe.ru
Segmentation, maps, satellite, public dataset
Usability
License
Tags
7/10
Description
We collected 50 satellite images from maps with a resolution of 700×400 and marked them up into three classes: houses (building), roads (road) and background (background)
The use of the dataset for commercial purposes is allowed only with the consent of the copyright holder (manager@labelme.ru )
The use of the dataset for commercial purposes is allowed only with the consent of the copyright holder (manager@labelme.ru )
View details
Such a dataset can be used to automate aerial surveys during cadastral work, as well as to improve the algorithms of navigators and maps. Screenshots of the map are taken from the YandexMaps service.
Color map:
background:0,0,0
building:115,51,128
road:32,224,224
The average segmentation time of one image is 4 minutes.
Color map:
background:0,0,0
building:115,51,128
road:32,224,224
The average segmentation time of one image is 4 minutes.
File dataset
Download a sample maps segmentation demo.rar
Good design is obvious. Great design is transparent.
Download the "maps segmentation" dataset from the portal Labelme.ru
Размер: (11.4 Гб)
Verified
Status:
33.1 Мb
Size:
Download the dataset "Segmentation of houses and roads on maps"
To download a dataset for free, enter your data
Free download of the Labelmen dataset