Mapflow - Get startedļ
Go to Mapflow and register or login using your email. You can also use Google account to login. Mapflow processes imagery and extracts vector objects from it. So as a user, you start processings, or as we call it, flows.
After registration, your first Default project will be created automatically and you will be redirected to the main dashboard:
Note
Note that in the Maplfow 2.0 update we joined all the platformās APIs, thatās why the projects migrated from Mapflow Web app are temporarily marked as WEB
:
Note
See also about the Team accounts to collaborate with other users
How to run your AI mapping in Web appļ
Mapflow is designed to be intuitive. Here is our step-by-step user guide:
1. Data sourceļ
1.1. Select the raster sourceļ
Imagery providersļ
Here you can select one of the default providers:
Mapbox Satellite is a provider of global high resolution satellite imagery. The date of the image and updates cannot be chosen.
ArcGIS World Imagery is a provider of the global coverage composed of the high and medium resolution satellite imagery and the aerial imagery, hosted by ESRI. The frequency of updating images is 3-5 years.
Hint
You can check the ArcGIS World Imagery metadata (date, resolution, and precision) by location. To do this, follow this link, zoom in, choose the point on the map and click.
To read more about satellite images check Satellite imagery
Custom URLļ
This tab allows you to:
Specify the XYZ link to the source of the images. The link is automatically checked for validity (for example, Openaerialmap -
https://apps.kontur.io/raster-tiler/oam/mosaic/{z}/{x}/{y}.png
);When adding you custom tile service it might be needed to define Y position: XYZ (top left), TMS/WMTS (bottom left);
Set source image coordinate reference system (espg:3857 or espg:3395);
Reset all entered custom parameters;
Upload GeoTIFFļ
Here you can upload your own image in GeoTiff format.
Warning
Currently, a preview of the uploaded image is not possible after loading the image, you will see only the area of its extent.
The processing AOI must be located in the area of this extent, otherwise the area will be cut off by the extent boundaries. The processing area is calculated by the intersection of the image extent and the AOI. If the AOI doesnāt intersect the image boundaries the error will pop up.
Image upload requirements (free plan):
The file size must be less than 1 Gb
Both sides image dimensions must not exceed 30.000x30.000 pixels
The image must be georeferenced and the CRS must be one of: - WGS84 (EPSG: 4326) - Web mercator (EPSG: 3857) - UTM (any zone)
Hint
If your image doesnāt meet the params above, we suggest to use Mapflow API / QGIS plugin which have more capabilities. Mapflow supports RGB imagery and also process single-band (panchromatic) imagery, but the AI models are not tuned for such kind of data, so the quality of the result may be worse than expected.
1.2. Specify the area of interestļ
This tab is used to add a processing area. The user can draw the area using Draw rectange / Draw polygon tool or download it in GeoJSON format (draw and download, as well as view the data structure, follow this link - geojson.io).
About Draw Polygon tool:
There may be confusion when drawing AOI using this tool, however, everything is quite simple. The area in this case is calculated not by the polygon, but by the Bounding Box (Bbox). Therefore, a warning MAX BBOX 100 KMĀ²
appears in the example below (bounding box is highlighted with yellow hatching):
Attention
Be aware that for now, only a single area can be drawn or uploaded per processing. If your GeoJSON file has multiple areas within its FeatureCollection, only the first one will be used. If you want to process multiple AOIs, you can split them into separate GeoJSON files and start processing for each one separately. Multiply AOI processing is supported in Mapflow ā QGIS and API and will become available in the Web app next releases. Other spatial data formats may also be supported for upload in the future, although we recommend using GeoJSON since it is a de-facto standard in a web mapping. It is natively supported by web mapping frameworks (e.g. Leaflet or Mapbox) and GIS like QGIS or the ArcGIS Suite.
2. AI modelļ
Select one of the AI models (see Model description).
3. Post-processing optionsļ
In this step, you can select additional processing options. Different AI models have their own options.
AI model āš Buildingsā:
Classification - we currently recognize the following building types: apartment buildings, single-household dwellings, industrial, commercial, other non-residential (see Urban Mapping classes).
Merge with OSM - compare the results with the buildings in OpenStreetMap and, if a sufficient overlap has been found, use the ones from the OSM instead.
Simplification - simplification and polygonization of building contours.
AI model āš² Forestā:
Heights - classifies vegetation by height: by default 0-4 m for shrub (ālow vegetationā), 4-10 m for regular forest (āmedium forestā), 10+ m for area with tall trees (āhigh forestā). Clasification is done per vegetated area, and not per single tree (see Forest Mapping classes).
4. Run the processingļ
At this step, you can set the name of your processing or use the initially generated one (the button allows you to generate variants of the name).
The āStart processingā will start processing.
The āClear Selectionā will clear the previously selected processing parameters.
Attention
Once you have selected the model and the processing parameters, you will see the total cost of your processing counted in Mapflow credits (our local currency units). Upon registration, you receive 250 credits for free for testing the platform (See the Mapflow credits).
View the resultsļ
In the āProcessing historyā you see the progress of processing as well as view and get the processing results. Each running processing displays the selected model, the status, the creation date and the processing cost.
View on the mapļ
After the successful processing completion, the card can be opened to view more detailed information about the processing parameters - AI Model, Post-processing, Area, Data Source.
Click the āView on the Mapā to quickly view the processing result on the built-in interactive map.
Feature attributesļ
Depending on the model and the options applied the extracted features might contain the semantic information that is written in the feature properties in GeoJSON.
E.g. āBuildingsā model with āsimplificationā and āclassificationā options:
PROPERTY |
DESCRIPTION |
---|---|
Area |
Feature area in meters |
Class_id |
Buildings typology (see the Urban Mapping classes) |
Shape_type |
The geometry primitive depending on the polygonization algorithm thatās been applied to the feature mask |
Processing_date |
The date of the processing completion |
Simplification_score |
The score is counted by the intersection of the pixel feature and polygonized feature. The higher score - the closer polygonized feature to the pixel one. |
Rate the processingļ
Important
We encourage you to rate the processing results. āļøāļøāļø Your assessment is important for further improvement of the Mapflow AI models. We review all the userās feedback and get back to you if any issues.
Interact with the processing resultsļ
The Processing menu includes the items as follows:
Click āDuplicateā on an existing processing to use its parameters as a starting point for a new one (on restart with the same parameters, new processing is started).
Click āSource detailsā on the existing processing to view information about source details and AOI.
Tip
If you want to duplicate processing to change some params, like choosing another AI mapping model for the same AOI - use Duplicate
Download GeoJSON.
Allows to download results as a file in GeoJSON format.
Warning
This option is not available in the free plan. Only paying users (see pricing) can use this option.
āOpen with geojson.ioā - view the results in the browser using geojson.io.
Tip
The geosjon.io app allows you to save the results in a different GIS-compatible format (CSV, KML, TopoJSON, WKT, Shapefile). Just click āSaveā and select the format of your preference in the menu bar.
āOpen with kepler.glā - view the results using this embedded app. It is a simple but powerful tool for geospatial data visualization and analysis.
You can find detailed information about Kepler.gl in their user manual.
Working with APIļ
Mapflow provides a REST API which, for example, allows you to query for currently running flows and fetch the results. If you are developing an application and want to use our API, - check out Mapflow processing API.
Important
You must follow the requirements specified with Model requirements when uploading your own images for processing through the API of the Mapflow platform. Send a request using data preprocessing to help@geoalert.io.