Abstract
Remote sensing requires fast and accurate analysis of remotely sensed images. However, the high demand for computation power has limited its important applications in real-time environments. This article describes a system integration approach to achieve real-time classification of satellite images by parallelism. In contrast to the traditional systems, which deal with data acquisition, compression, transmission, and analysis separately, our smart remote sensing system integrates a pipelined architecture onboard a satellite and a network of workstation clusters at the ground station. The pipelined system is responsible for the enhancement and compression of the image data captured from the camera onboard a satellite. A network of workstation clusters at the ground station is dedicated to the comprehensive analysis of such preprocessed remote data transmitted from the satellite in a parallel virtual machine environment, which includes decoding of the compressed image data and image classification by textures. Both the system design and implementation strategies are briefly described. In addition, a parallel search algorithm is introduced to speed up the classification tasks based on the accelerated cascading technique and the dynamic processor allocation scheme. The time complexity analysis and experimental results show the effectiveness and efficiency of the proposed techniques.
Original language | English |
---|---|
Pages (from-to) | 138-148 |
Number of pages | 11 |
Journal | International Journal of Robotics and Automation |
Volume | 18 |
Issue number | 3 |
Publication status | Published - 1 Oct 2003 |
Keywords
- Feature extraction
- Parallel computing
- Parallel virtual machine (PVM)
- Pipeline architecture
- Real-time remote sensing
- Texture analysis image classification
ASJC Scopus subject areas
- Software
- Control and Systems Engineering
- Modelling and Simulation
- Mechanical Engineering
- Electrical and Electronic Engineering
- Artificial Intelligence