DMD-Based Hadamard Transform Near-Infrared Spectrometer Design and Implementation of Fast Processing System
WANG Shuo1, 2, XIE Zhen-kun1, 2, WEI Zhi-peng1*
1. State Key Laboratory of High Power Semiconductor Laser, Changchun University of Science and Technology, Changchun 130022, China
2. Zhongshan Institute of Changchun University of Science and Technology, Zhongshan 528437, China
Abstract：The near-infrared spectrometer based on a digital micromirror device (DMD) has the advantages of good wavelength repeatability, high resolution, and good vibration resistance. It is widely used in the fields of food safety and agricultural production. With the development of micro near-infrared spectrometers based on DMD becoming more and more mature, the cost and performance of the instrument are still the key to research and development. Although most researchers have focused on software development and detection methods, the processing speed of the instrument hardware is also crucial. The spectral analysis can be effectively realized only by ensuring that the spectrometer can collect and transmit data quickly and accurately. In addition, in most studies, the generation and decoding of the Hadamard matrix is usually completed by the host computer, and the template is imported by FLASH storage. However, this method may limit the time efficiency of collecting complete spectral data. A method of high-speed driving DMD and fast data acquisition is proposed to improve the spectral acquisition speed and signal-to-noise ratio. A hardware circuit system is designed based on a DMD miniature near-infrared spectrometer. The system adopts the architecture of Field Programmable Gate Array (FPGA) and ARM and innovatively realizes the generation and decoding process of odd-even Hadamard template at the bottom of the embedded system, which accelerates the speed of spectral analysis and improves the signal-to-noise ratio. By comparing with the DMD spectrometer on the market, the research results show that the spectrometer developed in this paper only takes 214 ms to complete the acquisition time of a single spectrum, which is 4 times higher than that of the commercial DMD spectrometer. In the same 3 s acquisition time, the signal-to-noise ratio of the spectrometer developed in this paper is 4 600, which is 1.5 times higher than that of the commercial DMD spectrometer. Furthermore, the spectral scanning of rapeseed samples was carried out by the spectrometer developed in this paper. The contents of fat, protein, and moisture in rapeseed were analyzed, and the corresponding models were established by partial least squares regression (PLSR). The correction correlation coefficient of rapeseed fat content was 0.986 5, and the prediction correlation coefficient was 0.967 2. The protein content correction correlation coefficient was 0.985 4, and the prediction correlation coefficient was 0.963 6. The correction correlation coefficient of moisture content was 0.987 5, and the prediction correlation coefficient was 0.961 4. The model evaluation results show that the spectrometer can meet the needs of rapeseed component detection and verify that the spectrometer has important application value in the commercial field.
[1] CHU Xiao-li(褚小立). Practical Manual of Near Infrared Spectroscopy Analysis Technology(近红外光谱分析技术实用手册). Beijing: Machinery Industry Press(北京: 机械工业出版社), 2016. 1.
[2] Bec K B, Grabska J, Huck C W. Chemistry—A European Journal, 2021, 27(5): 1514.
[3] YU Fan, WEN Quan, LEI Hong-jie, et al(庾 繁, 温 泉, 雷宏杰, 等). Laser & Optoelectronics Progress(激光与光电子学进展), 2018, 55(10): 100003.
[4] WANG Su-hui, ZHANG Xu, ZHANG Gen-wei, et al(王宿慧, 张 旭, 张根伟, 等). Infrared Technology(红外技术), 2020, 42(7): 688.
[5] HUO Xue-song, CHEN Pu, DAI Jia-wei, et al(霍学松, 陈 瀑, 戴嘉伟, 等). Journal of Instrumental Analysis(分析测试学报), 2022, 41(9): 1301.
[6] DeVerse R A, Hammaker R M, Fateley W G. Applied Spectroscopy, 2000, 54(12): 1751.
[7] Xiang D, Arnold M A. Applied Spectroscopy, 2011, 65(10): 1170.
[8] Xu J, Liu H, Lin C, et al. Optics Communications, 2017, 383: 250.
[9] Lu Z, Zhang J, Liu H, et al. Micromachines, 2019, 10(2): 149.
[10] Chen X, Quan X. Sensors, 2022, 22(16): 6237.
[11] FENG Hai-zhi, LI Long, WANG Dong, et al(冯海智, 李 龙, 王 冬, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2023, 43(1): 16.
