Page 1 of 3 Next  Asheer Kasar Bachoo ICT4E0, Meraka Institute, CSIR, South Africa abachoo@csir.co.za Frederic Claudel ICT4E0, Meraka Institute, CSIR, South Africa frederic.claudel@magellium.fr Open source software (OSS) has been present for a number of years. In its infancy, it may have been viewed as an overly ambitious sphere that has greatly increased the throughput of GISs. In this article, we discuss open source software used at the Meraka Institute's Information and Communications Technology for Earth Observation (ICT4E0) Unit for developing a high speed satellite image experimental project. But since then, the markers have shifted considerably and organizations can now effectively develop software applications for their own specific requirements using OSS. One sector that can benefit greatly from the open source concept is that of GISs and remotely sensed data. A concrete example of OSS for remote sensing applications is in the field of digital image processing and high performance computing, a registration application. Open source software (OSS) Open source applications are those distributed with their source. Their licenses afford users the privilege to use and modify the program for any purpose, as opposed to closed (proprietary) software. These modified programs can then be redistributed without paying royalties to previous developers. An important component of open source software is the GNU public license (GPL) [1]. It ensures that a particular piece of software is free for all users and it protects the user's rights. In the context of open source, the word "free" means freedom - users can modify, distribute and use it in other free software. In addition, distributors may also charge a fee for the program. An advantage of free software is that its evolution into something powerful is rapid due to it being used and developed by a large contingent of users. There are a number of people that are not willing to embrace OSS - they rely on proven systems for their productivity and development. OSS needs to be integrated into these environments over a reasonable time frame, just as proprietary software develops and integrates into a system. This will ensure a seamless process. The power of OSS software lies in controlling the core of your software products and understanding its finer workings for optimal productivity and development. The association of "free" software with poor quality is an unfounded concept. High performance computing High performance computing can be performed by specialized hardware, software or a combination of both. In open source spheres, software is available for cluster computing. A cluster is a computing environment consisting of many computers connected to each other by a local area network. This network can be viewed as a single machine with large processing capabilities. It is also a cost-effective and powerful option for high speed computing, known as a commercial off the shelf (COTS) system. COTS systems can be tailored for specific applications and provide an easy way to facilitate expansion of an existing system at a very low cost. An automatic satellite image registration prototype The CSIR's ICT4E0 research unit has developed OSS modules for satellite image registration. These programs use and expand the Open Source Software Image Map (OSSIM) library, written in C++, to perform accurate image processing of remotely sensed data [2]. The OSSIM library supports a wide range of data formats, projection and sensor types and image processing algorithms. It also incorporates a Geospatial Data Abstraction Library (GDAL) plug-in for wider access to a range of data formats. The registration toolkit consists of three modules - correl, modopt and rejout - that perform image correlation, model optimization and outlier rejection respectively. A brief explanation of each module is provided in the following subsections.
Page 1 of 3
Next |