摘要
随着DSP处理能力的不断增强以及图像处理算法的日益完善,开发基于机器视觉的嵌入式实时系统成为目前图像处理领域的一项重要分支。为此,介绍了基于TMS320DM642的嵌入式棉籽识别系统的硬件原理和软件结构。首先分析了DM642的结构和特点,然后阐述了系统的组成及关键技术,最后介绍了基于RF5参考架构的软件设计方案。经验证,本系统能够处理两路视频输入,实时地完成棉籽识别。
With the development of the DSP processing ability and the perfection of the image processing algorithm increasingly, it has become an important branch in image process to develop embedded system in real-time based on machine vision, in this paper, it introduced the hardware theory and software structure of the cottonseed identifying system which is based on TMS320DM642. It analyzed the structure and characters of DM642 first, and then the expatiated system＇s constitution and the key technology. Finally explained the software design scheme based on RF5. By validation, the system can process two input video and finish identifying cottonseed in real-time.
出处
《农机化研究》
北大核心
2007年第2期178-181,共4页
Journal of Agricultural Mechanization Research
关键词
农艺学
嵌入式棉籽识别系统
设计
DSP
DM642
智能识别
agronomy
embedded cottonseed identification system
design
DSP
DM642
intelligent identification.
