ActivitiesSoftware OfferContacts

 

IDRISI Image Processing

 

The final component of the TerrSet constellation is the IDRISI Image Processing toolset. The raster foundation of TerrSet provides an exceptional base for the analysis of remotely sensed imagery. Tools can be broadly classified into restoration, enhancement, transformation and classification. However, while the system provides a broad-spectrum approach to the analysis of remotely sensed images, the emphasis is unquestionably on the conversion of remotely sensed images to interpreted maps. The tools for classification are the broadest in the industry, including both hard and soft classification procedures based on machine learning (such as neural networks) and statistical characterization. Elements include:

A complete set of geometric and radiometric correction tools including resampling, mosaicking, destriping and atmospheric correction. A special procedure is provided for Landsat import (including Landsat 8).

Image enhancement tools include stretching, compositing, filtering and sharpening images. For systems that include a higher-resolution panchromatic band, pan-sharpening procedures include a modified IHS (Intensity-Hue-Saturation) or Hyperspherical Color Sharpening (HCS) procedure. Spatial Fourier Analysis provides special opportunities for the removal of noise.

TerrSet’s broad set of raster tools provides unlimited possibilities for transformation. However, special facilities exist for Principal Components and Canonical Correlation Analysis and vegetation index creation.

TerrSet contains the broadest set of classifiers (procedures for the computer-assisted interpretation) for remotely sensed imagery in the industry. Supervised classifiers include parallelepiped, minimum distance, maximum likelihood, Fisher Linear Discriminant Analysis, and K-nearest neighbor (KNN). Unsupervised classifiers include ISODATA, chain clustering, histogram peak and K-Means. Machine-learning classifiers include Classification Tree Analysis (CTA) and four neural network classifiers: Multi-Layer Perceptron (MLP), Self- Organizing Map (SOM), Fuzzy ARTMAP, and Radial Basis Function (RBF). Soft classifiers and mixture analysis tools include a Bayesian probability classifier, a Mahalanobis distance classifier, fuzzy set classifier, and linear spectral unmixing. Hyperspectral image analysis and classification is included. A special set of procedures are provided for segment-based classification. Tools for accuracy assessment are also provided.

IDRISI Image Processing

TerrSet includes the IDRISI Image Processing toolset with the most extensive set of classifiers for remotely sensed imagery in the industry. In this example, a Landsat 8 image of Rhode Island for May 2103 was classified using a Maximum Likelihood classifier and training data for classes consistent with the U.S. National Land Cover Database (NLCD).