ERDAS

ERDAS (Earth Resources Data Analysis Systems) has been a major provider of software for multispectral image analysis integrated with raster geographic information system (GIS) functionality since the early 1980s. In May 2001, ERDAS was acquired by Leica Geosystems of Switzerland as a part of an effort to broaden its geoprocessing capabilities. ERDAS IMAGINE is now a broad collection of software tools designed specifically to process imagery that exists within the Leica Photogrammetry Suite (LPS).

ERDAS was a spin-off of research being performed at the Georgia Tech Engineering Experiment Station (EES), now called the “Georgia Tech Research Institute.” EES in the early 1970s developed public domain image processing software for NASA and performed a statewide land cover classification of NASA's Earth Resources Technology Satellite (ERTS) data. The land cover maps were analyzed by county and watershed, with area coverage being calculated for each unit.

Building on their experience gained in the development of software for the computer processing of multispectral images and in the application of early GIS, such as IMGRID developed by the Harvard School of Design, the founders initially formed ERDAS in 1978 as a consulting company whose mission was to provide services in environmental analysis of satellite and other spatial data sets. ERDAS developed software for digital pattern recognition using multispectral satellite data from ERTS and for integration of other spatial [Page 128]databases (soils, elevation, slope, etc.) with the land cover information derived from the satellite data.

ERDAS initially developed its consulting software on a Data General 16-bit minicomputer using algorithms derived from the literature. Experience at Georgia Tech in the implementation of complex algorithms in the limited environment of minicomputers allowed the restructuring of the mostly mainframe-based image processing and geographic database analysis tools into an interactive set of software that could easily be used for project-oriented consulting.

From 1978 to 1980, consulting projects for ERTS (now Landsat) analysis and the development of geographic raster databases for large-area planning were the mainstays of ERDAS. In addition, a mobile version of the minicomputer system was developed for NASA Goddard to provide image processing capability in their mobile van. During this time, repeat customers for the land cover analysis and geographic database integration began to ask ERDAS for a software/hardware system so that they could do their own satellite image analysis. There were several minicomputer-based, commercial image processing systems currently on the market, including the Image 100 from General Electric and systems from ESL and I2S. These systems were very expensive ($500,000 to $1,500,000) and not within the range of most potential users other than government agencies and oil companies. ERDAS began to investigate what it would take to create a system that would be affordable and easy to use and yet provide the same functionality as the larger, more expensive systems.

In 1979 and 1980, hobby microcomputers such as Altair, Motorola, Cromemco, and so on, were becoming popular, and several students working at ERDAS were involved with the trend. Because of the modularity and line-by-line access to images and databases, ERDAS became interested in what could be done in terms of real analysis on microcomputers. A challenge was put to the students to try to implement some of the simple image processing and GIS algorithms on a hobby computer. Even though the microcomputers had very little memory and limited access to disk storage, the efficiency of the raster implementation made it possible to implement most algorithms on the microcomputers.

...