A new locally-adaptive classification method LAGMA for large-scale land cover mapping using remote-sensing data

Abstract

A new locally-adaptive image classification method LAGMA (Locally-Adaptive Global Mapping Algorithm) has been developed to meet requirements of land cover mapping over large areas using remote-sensing data. The LAGMA involves the grid-based supervised image classification using classes’ features estimated locally in classified pixels’ surrounding from spatially distributed reference data. The LAGMA considers inherently spatial variations of classes’ features and is capable of exploiting discriminative properties of local classes’ signatures without any preliminary stratification of mapping area. The LAGMA has been applied for country-wide land cover classification over Russian Federation using the Vegetation instrument data on board of the SPOT (Satellite Pour l’Observation de la Terre) satellite and has demonstrated advantages in terms of recognition accuracy.

Acknowledgements

The authors wish to thank Dr. Craig Cassells, Editorial Assistant for Remote Sensing Letters, and two anonymous reviewers for their valuable comments and suggestions to improve this article.