用于食品安全、 环境污染、 毒品、 化学战剂检测等的现场快速检测要求检测设备便携、 快速和准确。 目前的常规实验室检测方法虽然能够实现准确检测, 但是其实时性较差, 无法满足现场快速检测的要求。 因此, 设计开发了基于等离激元增强拉曼光谱的便携式快速检测仪, 能够实现对滥用添加剂、 违禁食品添加剂、 农药残余、 毒品、 化学战剂和环境水污染等大量有害物质的现场快速定性检测。 该检测仪基于ARM嵌入式系统开发, 编写了其操作交互界面和底层驱动程序, 实现了拉曼光谱数据自动标定和特征谱图快速识别等算法。 在仪器中搭建了检测物质拉曼光谱的标谱数据库, 数据库根据检测科目进行分类, 每种科目包含该类别的多种物质, 每种物质均包含其高中低多种浓度的标准品和样本谱图。 设计实现了大类科目的辨识比对算法和GPU硬件加速算法, 实现了对某一类科目的所有物质的快速比对, 与传统的纯CPU算法实现相比较, 在辨识速度上提高了20倍以上。 通过市场购买的功能饮料、 违规葡萄酒与果汁等实际样品对所制备的样机仪器进行测试, 测试结果符合预期, 具有良好的灵敏度与重现性, 满足了现场快速检测的要求。

The detection equipment for food safety, environmental pollution, drug safety, chemical warfare agents requires portable, fast and accurate detection. The traditional laboratory test methods can meet accurate requirement, but it is slow and cumbersome. This paper proposes a portable and rapid detection equipment based on plasmon-enhanced Raman spectroscopy, which can realize rapid qualitative and accuracy detection. The equipment system is based on ARM embedded architecture. The UI interface and drivers are finished and the algorithms of auto calibration of Raman spectroscopy data and characteristic spectra fast recognition is designed. The spectrum database of standard subjects is constructed in this local equipment, which is classified according to the test required subject. The database includes a variety of substances in every subject, which contains standard and sample spectrum of different concentration. The algorithms of multiple substances identification and GPU hardware acceleration are designed and presented in this paper, which can detect mixed sample with more than twenty times faster then CPU soft calculate. A series of experiments are finished in the end to verify the algorithms and performance of the proposed equipment, such as wine, fruit juice and functional drinks getting from shop. The result meets the expectation with good sensitivity and reproducibility.