基于FPGA的成像光谱仪多通道实时光谱识别设计

实时光谱识别是目前干涉式成像光谱仪领域研究的热点问题之一,其难点在于复原光谱数据量大及对识别算法的实时性要求高。基于FPGA器件设计了一种基于光谱角匹配(SAM)算法的干涉式成像光谱仪多通道实时光谱识别系统。该系统将多个不同的参考光谱分别存储在其对应通道的存储器中,根据SAM算法实时计算输入光谱与各通道参考光谱之间的光谱角,通过对各通道输出的光谱角进行比较而实现光谱的实时识别。仿真结果表明,系统在Xilinx芯片XC7A100TFGG484-2上对复原光谱进行实时有效识别的运行频率可达100MHz。系统以流水线方式运行,具有速度快、体积小、便于升级等优点,为干涉式成像光谱仪实时光谱识别的工程实现提供了一种有效的技术途径。

Design of Multi-channel Real-time Spectral Recognition for Imaging Spectrometer Based on FPGA

Real-time spectral recognition is one of the hot issues in the field of interferometric imaging spectrometers. The difficulty lies in the large amount of recovered spectral data and the high real-time requirements for recognition algorithms. In this paper, a multi-channel real-time spectral recognition system based on spectral angle matching (SAM) algorithm for interferometric imaging spectrometer is designed based on FPGA device. The system stores a plurality of different reference spectra in the memory of its corresponding channel, calculates the spectral angle between the input spectrum and the reference spectrum of each channel in real time according to the SAM algorithm, and compares the spectral angles of the output of each channel to identify spectra in real time. Simulation results show that the system can effectively identify the restored spectrum on the Xilinx chip XC7A100TFGG484-2 in real time up to 100MHz. The system runs in a pipeline mode, which has the advantages of high speed, small size and easy upgrade. It provides an effective technical approach for the real-time spectral recognition of interferometric imaging spectrometers.