The core of the Mapflow are the Mapping Models. Mapflow enables to detect and extract features in satellite and aerial images powered by semantic segmentation and other deep learning techniques.

AI-Mapping Models

🏠 Buildings

Extracting of roofprints of buildings from imagery of high resolution.

Additional options:

  • Classification by types of buildings – typology of buildings is represented by the main classes (see reference).

  • Building heights - building height estimation by the length of the shadow and the visible part of the wall. Shift to the building footprint.

  • Simplification - The algorithm allows you to correct the irregularities of the contours of our model.

  • Merge with OSM - This option allows you to replace the obtained data of our model with data from the OSM, if the polygons of the OSM buildings and the model overlap significantly (Jaccard coefficients - more than 0.7).

🏙️ High-density housing

Our “high-density housing” AI model is designed for areas with terraced or otherwise densely built buildings, common in the Middle East, parts of Africa, etc.

🎄 Forest

Extracting the forest masks from RGB images of high resolution (2 meters) without classification by type, density and heights.

Additional options:

  • Classification by heights – classification the areas of vegetation and shrub vegetation by height classes according to the specified thresholds: 0-4 m, 4-10 m, 10+ m, and division of classes in 4+ m according to the density of vegetation into: dense and sparse. Forest areas of each height class are polygonized in separate features. The height class and density of vegetation are specified in the polygon properties.

🚗 Roads

Extracting the road mask from satellite images of high spatial resolution.

🏗️ Construction

Detection of the construction sites by classification of tiles of hi-resolution satellite images.

Models reference

Buildings

Buildings

Description

Model input, min. GSD m/px

Segmentation

Extract roof contours (roofprints) from high-resolution satellite imagery

RGB 0.5

Classification

Here are the types that we currently recognize: apartment buildings; single-household dwellings; industrial; commercial; other non-residential

RGB 0.5

Building heights

For each building, we estimate its height using its wall’s and shadow’s lengths. If height detection option is selected, all roof contours are shifted accordingly, i.e. converted to footprints

RGB 0.5

Merge with OSM

Replacing the outlines of buildings of our model with the outlines from OSM

RGB 0.5

Simplifiation

Conturs correction of buildings

RGB 0.5

Forest

Forest

Description

Model input, min. GSD m/px

Segmentation

Extract segmentation masks of forested areas from high-resolution RGB images

RGB, 2

Classification

Classify the areas of vegetation and shrub vegetation by height and vegetation density.

RGB, 0.5

Roads

Roads

Description

Model input, min. GSD m/px

Segmentation

Extract roads from high-resolution satellite imagery

RGB, 1

High-density housing

High-density housing

Description

Input data, min. GSD m / pix.

Segmentation

Extraction of the contours of buildings with dense buildings on the roofs of buildings on satellite images of high spatial resolution

RGB 0.5