论文部分内容阅读
为了解决红外点目标检测算法的复杂性与系统实时性之间的矛盾,设计了一套基于DSP的实时处理硬件实施方案。首先,介绍了基于高性能DSP的实时处理系统的原理结构,在此基础上,从硬件和软件两方面提出了系统实时实现的关键技术,包括大容量图像数据的双缓冲交互访问机制、图像预处理算法的FPGA快速实现,以及基于多媒体指令集等技术的代码级优化策略。实践结果证明,优化后的实时处理实现方案有效地提高了系统的实时性和稳定性,能够满足红外点目标检测算法大数据量高速传输、复杂运算的实际需求。
In order to solve the contradiction between the complexity of the infrared point detection algorithm and the real-time of the system, a real-time processing hardware implementation scheme based on DSP is designed. First of all, the principle structure of real-time processing system based on high-performance DSP is introduced. Based on this, the key technologies of real-time system implementation are put forward in hardware and software, including the double buffering and mutual access mechanism of large-capacity image data, FPGA processing algorithm fast implementation, and based on multimedia instruction set technology such as code-level optimization strategy. The practice results show that the optimized real-time processing scheme can effectively improve the real-time performance and stability of the system, and can meet the actual demand of high-speed data transmission and complex calculation of infrared point target detection algorithm.