Interactive Digital Image Processing Investigation.
IBM FEDERAL SYSTEMS DIV GAITHERSBURG MD
Pagination or Media Count:
The objective of this investigation was to develop an interactive software package of general utility that can be used to support experiments and evaluate a class of interactiveautomatic digital techniques whose goal is feature extractionexploitation. The initial task was to survey and select techniques, algorithms and supporting functions that were to be included in the system. A maximum likelihood supervised classification algorithm, and an unsupervised classification clustering method. These and the associated support and evaluation functions were developed and implemented on the Digital Image Analysis Laboratory. Testing of the system was accomplished by performing experiments on areas for which both multi-dimensional images and ground truth were available. Analysis of experimental results validated the system. These experimental results were in good agreement with ground truth and further, supervised and unsupervised classification results were in substantial agreement. This report describes the data processing algorithmstechniques, software system, user guide information, and the experiment procedures used along with their results.
- Human Factors Engineering and Man Machine Systems