基于DSP的高速多光谱探测系统研究

为了实现高速获取多窄脉冲光谱分布数据信息,并抑制背景噪声及多光谱源之间相互干扰的问题,研究了一种基于DSP控制DMA高速采集干涉条纹并反演光谱的系统.设计了帧减降噪处理算法用于完成对窄脉冲光信号与背景光噪声的分离与提取,并通过3级兵乓缓存结构实现了该算法在高速DSP中的流水线作用.给出了系统硬件结构设计与DMA控制方式,分析了帧减降噪算法及兵乓缓存设计的实现方法.在CCS环境下仿真计算了帧减降噪算法条件下干涉条纹的输出变化,论证了该方法的可行性.实验采用555 nm、665 nm以及805 nm 3种窄脉冲激光同时入射系统进行光谱测试,数据采集由DSP5510控制DMA8237-5实现.结果显示,采用帧减降噪算法的光谱分布旁瓣抑制效果好,背景噪声基本消除,并且其处理速度比基于C++光谱分析软件的系统提高了3倍左右.由此可见,该设计具有高速多光谱同时探测的性能.

The Study on Multispectral Detection System Based on High-Speed DSP

In order to achieve high-speed acquisition of multi-narrow pulse spectral distribution data,and suppress the problem of mutual interference between background noise and multispectral sources,the system for inversion spectra based on DSP is studied,which can control DMA high-speed acquisition of interference fringes.The frame noise reduction algorithm is designed,which is used to complete the separation and extraction of narrow pulse light signals and background light noise.At the same time,in the high-speed DSP,the algorithm's pipeline function is implemented through the three-level buffer structure.The system hardware structure design and the DMA control mode are given,the frame reduction noise reduction algorithm and the military bucket cache design means are analyzed.In the CCS environment,the output variation of the interference fringes based on the frame noise reduction algorithm is simulated and the feasibility of the method is proved.In the experiment,three narrow pulse lasers,which are 555 nm,665 nm and 805 nm,are used for the incident system for spectral testing.Data acquisition is controlled by DSP5510 DMA8237-5 implementation.The results show that the spectral distribution side lobe suppression effect is good,the background noise is basically eliminated by the frame noise reduction algorithm.And its processing speed increases by about three times than the C++ spectrum analysis software based on the system.It can be seen that the design is a high-speed multi-spectral simultaneous detection performance.