GSI’s unique forest service is based on an analytical framework (proprietary and patented algorithms) applied to the source material available through satellite (and other) data. The process starts by asking the client three questions:
a). What data/information is required?
b). Relating to what area or region?
c). What data is currently available?
The answers to these questions then lead to a design for the software development process as follows:
- Concept development
Development of novel concepts for the processing, analysis and interpretation of satellite data applicable to a wide range of forestry and natural resource information needs as indicated by the client using the proprietary algorithms developed by the GSI team.
- Making new concepts operational
Implementation of new concepts for automatic operation in a high-performance environment, so that the algorithms can run on large datasets in short timescales.
- Machine Learning (applied on a per-client basis)
Using our machine learning software to calculate a range of required characteristics for the client’s area(s) of interest. This uses our optimised implementations of the algorithms and exploits the high performance super computer facility at EPCC (University of Edinburgh) which is capable of rapidly processing very large amounts of data.
- Data Analysis and Quality Assessment/Validation of Results (applied on a per-client basis)
As part of the project completion process, perform analysis of results generated, and present results in client specific format. At time we iterate machine learning with alternative data/configurations, when useful to improve results.
In each case, the GSI process differs and is customised to the client’s requirements.
The information output of the process will provide detailed maps and statistics for client-defined regions covering a range of forestry-applicable data, such as: extent of canopy cover, amounts of merchantable timber, total carbon content, etc.
Sample the GSI process here:-