基于改进的BSP模型的红外图像并行处理研究

针对导弹末制导红外图像处理实时性要求,分析了红外图像并行实时处理算法及其集群计算机的实现,建立了适应于实时处理的BSP改进模型和集群并行处理软件系统,并给出了红外图像并行处理算法的实现方法.研究了应用提升小波实现图像并行处理算法及实现,进而分析了算法性能,推导出了加速比模型.实验验证了其有效性.

The Study on Parallel Computing of Infrared Image Based on BSP Computing Model

Considering the real-time requirement of processing infrared images on missile flight in end-game,this paper makes an analysis on parallel algorithms for the real-time processing of infrared images and on the realization of cluster computers.The BSP algorithm model for real-time processing and the cluster software system for parallel processing have been set up,with the suggested means of realization of parallel algorithms for processing infrared images.Meanwhile,it also makes a research on the parallel algorithms by application of Lifting Wavelet Transform as well as.their realization,analyzing hence the performance of the Lifting Scheme with derivation of the speedup-ratio model.In addition,the model's validity has been verified by experiments.