论文部分内容阅读
针对各个滤波算法的适用范围不同,仅仅依靠单一的滤波算法不能适应红外系统的实际需求,提出了复杂背景下红外图像弱小目标的多算法并行处理硬件结构,该结构采用了基于自适应“投票表决”的多源数据决策技术,能根据外场的实际情况自适应地将各种滤波算法的优势集合起来,可以极大地改善滤波效果,提高点目标的检测概率,为后续处理降低数据量及运算量。针对硬件仿真阶段注入原始红外图像数据流困难,无法实时对比图像处理前/后效果的问题,基于Wishbone-PCI桥核设计并实现了以FPGA为主处理器、主机为从处理器的红外图像处理仿真平台。实验表明,多源数据决策技术很好地满足了系统的实际需求,多算法并行处理硬件结构运行稳健有效,硬件仿真平台可靠稳定,整个系统有很强的工程实用价值。
According to the different application range of each filtering algorithm, a single filtering algorithm can not adapt to the actual demand of the infrared system. A multi-algorithm parallel processing hardware structure for the weak target of the infrared image under the complicated background is proposed. The structure adopts the “ Voting ”multi-source data decision-making technology, according to the actual situation of the field can adaptively combine the advantages of various filtering algorithms can greatly improve the filtering effect, improve the detection probability of point targets for the follow-up to reduce the amount of data and Computation. In view of the difficulty of injecting the original infrared image data flow in the hardware simulation stage, it is impossible to compare the effect of before / after image processing in real time. Based on the Wishbone-PCI bridge core, an FPGA-based processor is designed and the host is an infrared image processor simulation platform. Experiments show that the multi-source data decision-making technology can well meet the actual needs of the system, and the multi-algorithm parallel processing hardware structure runs steadily and effectively, the hardware simulation platform is reliable and stable, and the whole system has strong engineering practical value.