![]() The total area changed through manual editing was 204.6km2 which was 20% of the total map area. This involved carrying out manual knowledge-based corrections of misclassifications and adding classes that could not be detected from EO data. The first iteration of the habitat map was refined by JNCC in consultation with ecological experts in the TCI. The LMM uses Random Forests, an ensemble classifier that creates a number of classification and regression trees, each trained by a subset of training data with randomly sampled variables. The classification method chosen for the generation of the habitat map was the Living Map Method developed by Natural England. Image segmentation was carried out using Trimble eCognition version 9.2.1 and the multispectral segmentation algorithm was used. OBIA involves segmenting imagery by grouping pixels based on homogeneity criteria to form objects, then classifying those objects based on their spectral and geometric properties, their relationships to other objects and to contextual data. The habitat map was produced using object-based image analysis. Pieslēgties vai Reģistrēties to request accessīiota flora and/or fauna in natural environment ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |