在遥感图像处理算法研究过程中,集成和利用已有算法代码往往存在技术瓶颈,算法的研制也缺乏统一的验证平台.文章对遥感图像处理算法的特性和软件构件技术进行分析和研究,提出一种面向SAR图像处理和展示的多语言构件模型RSIC.基于多语言构件模型,设计了一种并行化的SAR图像处理与展示架构和工作流组装方案.在SAR图像处理与展示平台的实际研发实践中,该方法可适应多语言编程环境,综合利用多种工具包,继承和复用现有算法代码.并能有效减轻编程工作量,提升遥感图像处理相关算法研究进度.
In new algorithm research and software engineering application of remote sensing image processing, development of core algorithms often lacks the unified evaluation and verification platform, and integration and utilization of the existing algorithm code also have a lot of technical bottlenecks, which seriously restrict the research and application of the remote sensing products. According to the characteristics of remote sensing image processing algorithms, and analysis and research on software component technology, the remote sensing image component (RSIC) model of multi-programming language was proposed for synthetic aperture radar (SAR) image processing and presentation in this thesis. We described the component model of remote sensing image such as I/O, preprocessing, post-processing, visualization and parallel processing by the extended Backus-naur form (EBNF). Then, the SAR images processing and presentation architecture were designed by hybrid heterogeneous parallel based on Hadoop+ OpenMP+CUDA, and the workflow assembly program of remote sensing image processing and presentation were proposed by XML workflow definition language (XWFDL). In the platform of algorithm research and engineering practice for SAR image processing and presentation, this method can be effectively integrated the multi-programming such as C # , C+ +, Java and Matlab and other languages. It can inherit and reuse the existing algorithm code components, and comprehensive utilization of typical remote sensing software development tool kit. The researchers can focus on the improvement and innovation of the remote sensing image processing algorithms, which can effectively reduce the programming workload, and significantly improve the application development of remote sensing image processing software.