论文部分内容阅读
在焊缝自动跟踪系统中,为了有效可靠地对焊缝进行跟踪,必须对焊缝图像进行滤波预处理,从而提高图像信号的信噪比。本课题采用了一种利用硬件实现的中值滤波算法,利用软核的设计方法,采用NiosⅡ处理器的定制指令,把强实时软件算法或费时的软件计算作为定制指令,加入到NiosⅡ处理器中进行指令集中,并且在QuartusⅡ的集成开发环境下编写HDL代码,进行综合;利用QuartusⅡ内部的仿真器对设计做脉冲响应仿真和验证。结论是满足了实时图像预处理时对速度的高要求。本课题的创新点在于NiosⅡ应用于焊缝位置提取设计。
In the weld automatic tracking system, in order to effectively and reliably track the weld, the weld image must be preprocessed by filtering to improve the signal-to-noise ratio of the image signal. The subject uses a median filter algorithm using hardware, the use of soft-core design method, the use of custom instructions Nios Ⅱ processor, the strong real-time software algorithms or time-consuming software as a custom instruction, added to the Nios Ⅱ processor Carry on the instruction centralized, and write the HDL code under the integrated development environment of QuartusⅡ, carry on the synthesis; Make the impulse response simulation and verification to the design with the Quartus Ⅱ internal simulator. The conclusion is to meet the high speed of the real-time image preprocessing requirements. The innovation of this subject is that Nios Ⅱ is applied to the design of the position of the weld.