论文部分内容阅读
本文介绍的波前处理机是１．２ｍ望远镜自适应光学系统的关键部件之一，吞吐率高，延时小，是具有高速数据处理能力的实时处理系统。为满足系统的高速度性能，系统设计在实时性、并发性的基础上，将图象数据采集加入流水线，并对波前处理中自然存在的子任务作“去相关”’处理，建立了图象采集、斜率计算、波前复原间流水线操作。图象读出时间的利用及流水线的建立，大大减少了时延，降低了系统实时性要求。流水线处理与高度并行处理相结合，也是本处理机设计的一大特色。
The wavefront processor introduced in this paper is one of the key components of the 1.2m telescope adaptive optics system. It has high throughput and low latency and is a real-time processing system with high-speed data processing capability. In order to meet the high-speed performance of the system, the system design adds the image data acquisition to the pipeline on the basis of real-time and concurrency, and “decouples” the naturally occurring sub-tasks in the wavefront processing to establish the graph Like acquisition, slope calculation, wavefront recovery between the pipeline operation. The use of image readout time and the establishment of pipeline greatly reduce the time delay and reduce the system real-time requirements. Pipeline processing combined with a high degree of parallel processing, but also a major feature of this processor design.