Identification of Pork Parts Based on LIBS Technology Combined With PCA-SVM Machine Learning
XU Yu-ting1, SUN Hao-ran2, GAO Xun1*, GUO Kai-min3*, LIN Jing-quan1
1. College of Science, Changchun University of Science and Technology，Changchun 130022, China
2. College of Mechanical and Electrical Engineering, Changchun University of Technology，Changchun 130012, China
3. College of Physical Science and Technology, Baotou Normal University，Baotou 014030, China
Abstract：In recent years, laser-induced breakdown spectroscopy (LIBS) is gradually emerging to classify and identify biological organizations by combining them with algorithms. Due to each part of the pork similar spectral characteristics, it is difficult to achieve accurate identifications only through the analysis of the effect of spectral information, so in this paper studied pork from four different parts of the same individual and sliced and planished them, then applied LIBS technology on four parts, i.e., the organization, fillet, plum flower, and front legs. 100 specimens of each sample were collected and the spectrum analysis was conducted. A preliminary analysis of the spectrum was performed on Ca, Na, K and 6 lines. It was found that other tissues were difficult to distinguish except for the C—N tissue of plum flower with more fat content and higher C content than other tissues, so the Principal Component Analysis (PCA) on these 6 principal components was carried out. The cumulative contribution rate of PC1, PC2 and PC3 reached 95%. The Support Vector Machines (SVM) classification model was established by employing feature scores as the input source of SVM model, and the confusion matrix diagram of these samples got obtained. Through observation of the confusion matrix, the classification accuracy of each type of samples could be clearly distinguished. The results showed that the accuracy of the four samples was 96%, 98%, 97% and 100%, respectively, with an average accuracy surpassing 97%. The study proved that LIBS combined with PCA-SVM can be used as a fast identification method for different parts of pork tissues.
Key words：Laser-induced breakdown spectroscopy; Principal component analysis; Support Vector Machines; Pork tissue
许毓婷，孙浩然，高 勋，郭凯敏，林景全. 基于LIBS技术结合PCA-SVM机器学习对猪肉部位的识别研究[J]. 光谱学与光谱分析, 2021, 41(11): 3572-3576.
XU Yu-ting, SUN Hao-ran, GAO Xun, GUO Kai-min, LIN Jing-quan. Identification of Pork Parts Based on LIBS Technology Combined With PCA-SVM Machine Learning. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2021, 41(11): 3572-3576.
[1] YUAN Wei-li，LIU Chao，WANG Yong-song，et al(袁伟丽, 刘 超, 王永松, 等). Acta Laser Biology Sinica（激光生物学报），2018, 27(4): 320.
[2] Varricchio E, Russolillo M G, Maruccio L, et al. European Journal of Histochemistry, 2013, 57(1): 10.
[3] LUO Jia-qin, WANG Jia-qi, BU Deng-pan, et al（罗家琴, 王加启, 卜登攀，等). Scientia Agricultura Sinica(中国农业科学), 2008，(7): 2112.
[4] TANG Yu-long, GUO Zhou-yi(唐玉龙, 郭周义). Acta Laser Biology Sinica（激光生物学报）, 2004，13(5): 386.
[5] Tognoni E, Palleschi V, Corsi M, et al. Spectrochimica Acta Part B: Atomic Spectroscopy, 2002, 57(7): 1115.
[6] Yao M, Huang L, Zheng J，et al. Optics & Laser Technology, 2013, 52: 70.
[7] Leon Radziemski, David Cremers. Spectrochimica Acta Part B: Atomic Spectroscopy, 2013, 87：3.
[8] LI Jia-ming，CHU Ying-bo，ZHAO Nan，et al(李嘉铭, 褚应波, 赵 楠，等). Chinese Journal of Analytical Chemistry（分析化学）, 2016, 44(7): 1042.
[9] WANG Cai-hong，HUANG Lin，LIU Mu-hua，et al（王彩虹，黄 林，刘木华，等）. Chinese Journal of Analysis Laboratory（分析试验室），2017, 36(1): 32.
[10] Su Yuanhang, Lin Ruiyuan, Jay Kuo C C. Expert Systems with Applications，2019, 118: 355.
[11] WANG Hai-yan，LI Jian-hui，YANG Feng-lei(汪海燕, 黎建辉, 杨风雷). Application Research of Computers(计算机应用研究), 2014, 31(5): 1281.
[12] Cocianu C. Economic Computation and Economic Cybernetics Studies and Research，2013, 47(1): 41.
[13] ZHU Yi-ning，YANG Ping，YANG Xin-yan，et al(朱毅宁, 杨 平, 杨新艳，等). Chinese Journal of Analytical Chemistry（分析化学），2017, 45(3): 336.
