基于达芬奇技术的SMFT终端实时图像处理系统研制

Development of SMFT terminal real-time image processing system based on davinci technology

针对太阳磁场望远镜(SMFT)终端海量图像数据的传输与处理等问题,设计并实现了基于达芬奇技术的嵌入式实时图像处理系统。利用UML建模抽象出核心对象,描述整个系统的行为并指导系统的设计。系统硬件上围绕双核(ARM+DSP)达芬奇处理器DM6446进行了设计和实现;系统软件上对ARM端的Monta Vista Linux操作系统和DSP端的DSP/BIOS嵌入式实时操作系统进行了客制化裁减和BSP开发;用户软件上设计并实现了一系列图像处理和识别功能,包括CMOS图像传感器成像控制、图像预览、图像预处理、太阳磁图实时提取、太阳图像FITS文件的生成、文件格式转换和数据压缩等。良好地满足了(SMFT)终端实时采集与图像处理的要求,提高了望远镜的观测效率。

An embedded real-time image processing system is developed based on DaVinci Technology, which can deal with massive images data transmission and processing during solar observation. It is innovative to use the UML model to describe the entire system. And hardware was designed around the double-core （ARM＋ DSP） DaVinci processor DM6446; The MontaVista Linux operating system on the ARM side and the DSP/BIOS real-time operating system on the DSP side are customized,And the BSP for both are developed. A series of image processing algorithms are designed and implemented for the application software,including control of CMOS image sensor, image preview, image pre-processing, solar magnetic map real time extraction, solar FITS file generation, file format conversion and data compression. The proposed system can Preferably meet the demand of real-time image acquisition and processing of SMFT, and improves the effidency of solar observation.