Applications for industries
Dynland can benefit your company at every stage of the value chain – by obtaining and analysing data, providing integrated platforms for users, or applying recommendations in real life. Our software solutions have been used for:
- Classification of multi- and hyper-spectral EO imagery for EO data providers.
- Forestry management. Species classification and stand stratification.
- Wetland analysis for peat extraction companies (peat societies), governmental
institutions responsible for maintaining and monitoring wetlands.
- Smart Farming for distinguishing the reasons of low vegetation – the lack of
nutrients in the soil (thus apply fertilizer) or land specifics – wind, soil moisture
and type (thus don’t apply fertilizer because there is no potential).
How Dynland software interprets images
Clusters are organised into a multi-layer network that allows users to regroup and assign classes (e.g. combine birch and oak forests into a single group for deciduous trees).
Each set of classes can be chosen according to the needs of the project at hand.
This technology will help you automate customised project flows.
The Latvian State Forest Research Institute «Silava» (www.silava.lv) performed an independent evaluation of the classification results on the previous page. This classification was accomplished based solely on images analysed by Dynland software.
Head-to-head comparison with the other algorithms
When processing synthetic data with various distributions Dynland is the only software solution that clusters all data meaningfully, regardless of distribution patterns.
The ArcGIS algorithm confuses logged forest and arable land even when the image is divided into 200 clusters
Dynland differentiates easily between logged forest and arable land
SAGA ISODATA confuses forest with city buildings and various meadow greens even when the image is divided into 200 clusters
Dynland differentiates easily between forest and meadow greens
The K-means algorithm confuses buildings with forest even when the image is divided into 200 clusters
Dynland differentiates easily between buildings and forest