Page 141 - Digital Analysis of Remotely Sensed Imagery
P. 141

Image Pr ocessing Systems     111

               integrate new modules or construct metamodules that control exist-
               ing IDRISI modules via a scripting language such as Visual Basic,
               Delphi, or Visual C++. Even the menu system can be completely
               reconfigured. Over the years this system has undergone drastic
               expansion and improvement. It has evolved from a system excellent
               for teaching the principles of digital image analysis to a professional
               system suitable for practical production of image analysis results at
               the industry standard. Its wide use by government agencies, schools,
               research institutions, academia, and the private sector testifies to its
               popularity (Simonovic, 1997).
                   Originating from a desktop system, IDRISI used to have a limited
               capacity to handle large image files. Now this limitation has largely
               disappeared with the improved functionality of personal computers.
               However, other limitations persist, such as the legacy of the highly
               modulized structure in which each module performs a narrowly
               defined step. This design philosophy may be ideal for teaching the
               concept of image processing, but is ineffective for practical produc-
               tion because a number of modules are needed to perform a simple
               analytical task such as image classification. The second critical limita-
               tion of IDRISI is its lack of automation. Although the Macro Modeler
               is a kind of scripting language for speeding up the process, it is not
               applicable to all functions. Besides, processing speed is slow for cer-
               tain functions such as zoom on screen (Huber, 2000). Finally, IDRISI
               still lacks data preparation functions, such as its inability to orthorec-
               tify images or georeference data using sensor-specific models. In spite
               of its comprehensive range of image classifiers, it is not possible to
               classify images based on pixel spatial properties and using external
               knowledge.


          4.2 ERDAS Imagine
               ERDAS is one of the oldest and leading geoinformatic software
               companies. Its major product, Imagine, contains a suite of compre-
               hensive and sophisticated tools for digital analysis of remotely
               sensed data. The latest release, version 9.1, offers something for
               everyone: 3D feature extraction, terrain modeling, hyperspectral
               data analysis, publication of 3D interaction environments, in addi-
               tion to photogrammetric tools after the company was acquired by
               Leica. This version significantly improves the integration of remote
               sensing and photogrammetry with the addition of the Leica Photo-
               grammetry Suite (LPS). Both Imagine and LPS can be configured to
               work closely with Oracle 10g Spatial and/or ESRI’s ArcSDE (Cothren
               and Barnes, 2006). The revamped ERDAS Imagine is offered at three
               levels, Essentials, Advantage, and Professional. Imagine Essentials
               encompasses a set of powerful tools for manipulating geographic
               and imagery data, such as image georeferencing, visualization, and
               map output. Imagine Advantage extends the capabilities of Imagine
   136   137   138   139   140   141   142   143   144   145   146