| 摘要: |
| 传统星上系统开发与研制通常是基于特定的专用宇航器件，这使得其开发周期长且可移植性较差。针对该问题，提出了一种基于Xilinx ZYNQ UltraScale+平台的ARM+可编程逻辑的解决方案。该方案用Vivado配置FPGA硬件架构，采用SDK(Software Development Kit)配置ARM实现两者数据交互。在整个系统设计中，通过集成在芯片内部的高速串行收发器和Aurora协议作为载体进行高速片间传输；采用自定义帧协议保证传输可靠性及安全性；采用AXI-Stream接口使其具有各类算法即插即用的灵活性；最后为验证本系统可以支持各类算法进行星上实时图像处理，通过在逻辑侧添加 Sobel 边缘检测算法进行验证。测试结果表明该系统图像数据传输无误码，星上串行高速传输速率较高，系统总体传输延时较低，且系统具有较强的算法通用性。 |
| 关键词: 星上数据传输 异步架构 Aurora协议 DMA 实时图像处理 |
| DOI：10.20079/j.issn.1001-893x.220624001 |
|
| 基金项目: |
|
| Design and Validation of a High-speed Satellite Image Data Transmission and Processing System |
| BAI Yu,HU Xuan,GENG Shisong |
| (1.College of Electronic and Information Engineering,Shenyang Aerospace University,110136,China;2.Institute of Software,Chinese Academy of Sciences,100190,China) |
| Abstract: |
| Traditional on-board system development and exploitation is usually based on specific dedicated astronautical devices,which makes its development cycle long and portability poor.To address this problem,an ARM+ programmable logic solution based on the Xilinx ZYNQ UltraScale+ platform is proposed.The solution uses Vivado to configure the FPGA hardware architecture and Software Development Kit(SDK) to configure ARM for data interaction between the two.The high-speed serial transceiver integrated in the chip and the Aurora protocol are used as carriers for high-speed inter-chip transmission.A custom frame protocol is used to ensure transmission reliability and security.The AXI-Stream interface is used throughout the system design to provide the flexibility of plug-and-play for various algorithms.Finally,to verify that the system can support various algorithms for on-board real-time image processing,the Sobel edge detection algorithm is added to the logical side.The test results show that the system has high on-satellite serial high-speed transmission rate,no error code for image data transmission,low overall transmission delay,and strong algorithm versatility. |
| Key words: onboard data transmission heterogeneous Aurora protocol DMA real-time image processing |
