SegOptim
SegOptim is an R package for object-based image analysis (OBIA). It allows to run, compare and optimize multiple image segmentation algorithms in the context of supervised classification. It also allows to perform unsupervised classification with several different methods and compare them through internal clustering metrics.
For more details about installation and how to use the package go to the tutorial here. Check also the paper describing the package functionalities here.
Functionalities
Currently the package offers several functionalities, namely: - Run different image segmentation algorithms; - Populate image segments with aggregate statistics (using pre- and/or user-defined functions); - Perform object-based supervised classification with several different methods; - Evaluate classification performance for single- or multi-class problems; - Optimize image segmentation parameters using Genetic Algorithms (GA) and other methods; - Compare different algorithms based on optimized solutions; - Perform unsupervised classification with several methods and compare the results using internal clustering criteria.
Algorithms Supported
Image segmentation algorithms: - ArcGIS Mean-shift - GRASS GIS Region Growing - Orfeo ToolBox (OTB) Large-scale Mean-shift - RSGISLib Shepherd’s k-means - SAGA GIS Seeded Region Growing - TerraLib 5 Baatz-Schape Multi-resolution segmentation and Mean Region Growing
Supervised classification algorithms: - Flexible Discriminant Analysis (FDA) - Generalized Boosted Model (GBM) - K-nearest neighbor classifier (KNN) - Random Forest (RF) - Support Vector Machines (SVM)
Unsupervised classification algorithms: - CLARA (Clustering LARge Applications) - Hard competitive learning algorithm - K-means - Neural gas algorithm