Abstract

We here present a reference database and three land use maps produced in 2017 over the Reunion island using a machine learning based methodology. These maps are the result of a satellite image analysis performed using the Moringa land cover processing chain developed in our laboratory. The input dataset for map production consists of a single very high spatial resolution Pleiades images, a time series of Sentinel-2 and Landsat-8 images, a Digital Terrain Model (DTM) and the aforementioned reference database. The Moringa chain adopts an object based approach: the Pleiades image provides spatial accuracy with the delineation of land samples via a segmentation process, the time series provides information on landscape and vegetation dynamics, the DTM provides information on topography and the reference database provides annotated samples (6256 polygons) for the supervised classification process and the validation of the results. The three land use maps follow a hierarchical nomenclature ranging from 4 classes for the least detailed level to 34 classes for the most detailed one. The validation of these maps shows a good quality of the results with overall accuracy rates ranging from 86% to 97%. The maps are freely accessible and used by researchers, land managers (State services and local authorities) and also private companies.

Highlights

  • We here present a reference database and three land use maps produced in 2017 over the Reunion island using a machine learning based methodology

  • The input dataset for map production consists of a single very high spatial resolution Pleiades images, a time series of Sentinel-2 and Landsat-8 images, a Digital Terrain Model (DTM) and the aforementioned reference database

  • Value of the Data - the referenced land cover maps provide an unprecedented overview of the entire territory of Reunion Island, with a significant potential impact in various tasks related to land, agriculture and environmental monitoring. - These maps can be used in GIS to monitor changes in the territory and help managers make decisions about urbanization, natural and agricultural land management. - The reference database can be used by remote sensing specialists to assess new methods for land cover mapping and other classification algorithms. - All data provided is georeferenced and in vector format for use in GIS tools in future projects

Read more

Summary

Specifications Table

A reference database consisting of GIS vector dataset in ESRI shapefile format composed of 6256 polygons representative of the diversity of land use on Reunion Island Fig. 1); Three land use maps produced starting from a dataset including a Very High Spatial Resolution (VHRS) Pleiades image, a time series of HRS Sentinel-2 and Landsat-8 images and a digital terrain model These maps correspond to three levels of land use nomenclature (from 4 to 34 classes) and are distributed in vector format (shapefile). Each geometry corresponds to an object provided by the segmentation of the Pleiades image, attributed using reflectances, radiometric and textural indices computed on the different remote sensing images available plus topographic information (altitude and slope) Such objects are individually classified using a supervised classification algorithm trained using the reference database described above. The advantage of distributing geographical data with this system is that it is possible to visualize them, to use them directly on GIS software with web services (WFS: Web Feature Service or WMS: Web Map Service) or download them

Materials
Images
The Moringa processing chain
The variables used in the classification
Findings
Post classification
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call