Digital image processing systems and remote sensing†
Fuat İnce · International Journal of Remote Sensing · 1983
Digital image processing (DIP) has been a primary reason for the emergence of remote sensing as an effective means of natural resources assessment. Increasing complexity and decreasing cost of digital hardware enable fast manipulation of data for maximum information extraction, which would be quite difficult if not impossible with classical photographic interpretation. DIP is viewed in terms of a basic requirement, which is interactive (near real time) operation, and two general characteristics, which are high volume of data and repeated application of a sequence of computations to each element of data. Computational and operational requirements for a range of DIP functions are reviewed. These include enhancements, transforms, geometric and radiometric corrections, encoding and classification. The implications are discussed in terms of hardware requirements. While a general purpose computer interfaced with a video display offers the simplest DIP system, it takes the pipeline and parallel architectures of display processors, and to a certain extent, array processors to execute most DIP functions in near real time, as demanded by the requirement of interactive processing. Several display processors, produced commercially or in research environ ments, are reviewed with respect to their architecture and processing capabilities. Some trends and expected future developments are also discussed.