Tools \ iota2
EO processing chain with machine learning
Source codeMaturity : Maintained | Categories : Processing Chains (pipelines), Preprocessing, Features Extraction | License : | Producer : CESBIO
Overview
iota2 is a land-use map production line for large areas based on time series of Sentinel or Landsat images. In order to facilitate methodological contributions, iota2 is open-source. As such, the channel makes extensive use of the image processing applications provided by Orfeo ToolBox.
iota2’s main features are:
- Automatic installation of the chain and its dependencies
- Use of sensors of different types (optical/radar) to obtain better quality classifications. Understanding of different N2A product formats (THEIA + PEPS) to be able to classify as many geographical areas as possible.
- Users can provide exogenous data as well as primitive data.
- Possibility of spatially delineating learning zones.
- Automatic calculation of primitives allowing the help to determine the label.
- Merging classifications.
- Automatic generation of coefficients allowing the evaluation of the quality of the classification.
- Division into stages of the chain allows it to be launched only on very specific stages.
- Combined use of Orfeo ToolBox multi-threading and MPI speeds up calculations.
Usage/Documentation
iota2 provide a command line interface.
The chain is fully configurable using a configuration file.
Some examples are provided in the documentation.
Installation
A conda package is available for iota2 installation. The package provides all needed dependencies including OTB.
Resources
Tutorials
There is no available tutorial for this tool.