Land Cover Mapping

Clark Labs' technologies include an integrated GIS and image processing software system, an excellent choice for land cover mapping applications with remotely-sensed data. Tools are provided for image restoration, enhancement, classification and transformation. Special techniques are included for soft classification and hyperspectral image analysis.

Application areas include:
- Satellite image processing
- Inventory and baseline land resource mapping
- Land change and time series analysis
- Agricultural monitoring
- Natural resource monitoring
- Error assessment and uncertainty management

Analytical Examples: [click images to enlarge]


Unsupervised landcover classification is available in IDRISI through the modules CLUSTER and ISOCLUST. CLUSTER, one of the fastest histogram peak clustering routines available, uses raw imagery to produce an output image of spectral classes. Either fine or broad clusters may be defined.


IDRISI supports export to GEOTIFF as well as import for 8-bit, 16-bit, 24-bit, and 32-bit GEOTIFF formats, such as this QUICKBIRD image.


IDRISI supports a variety of image restoration and transformation tools. In this image, the module ATMOSC is used for atmospheric correction.


IDRISI includes a neural network classifier using back propagation. Users have complete control over all parameters, such as the number of hidden layers, the learning rate, and the acceptable RMS.


Techniques for Hyperspectral Image Analysis are available in IDRISI. Tools include absorption spectra analysis using continuum removal for estimation of the degree of support for members of a library of spectral response curves developed in a laboratory setting, an unsupervised classifier, and several supervised classifiers including orthogonal subspace projection and linear spectral unmixing.


Linear spectral unmixing is available in the module UNMIX. Three options are provided: the standard unmixing method, a probability guided method, and an exhaustive search method. The latter two methods allow the user to evaluate more classes than the standard unmixing method.