基于多标记深度森林算法的冷鲜羊肉新鲜度无损检测方法

doi:10.3964/j.issn.1000-0593(2024)02-0580-08

摘要
参考文献
相关文章 (15)

全文: PDF (4460 KB)
输出: BibTeX | EndNote (RIS)

摘要：羊肉新鲜程度受多种因素影响，其检测一般要从感官性状、分解的理化产物和微生物繁殖程度等方面进行。然而基于单一指标的羊肉新鲜度检测局限性大，适用性低，很难综合评价羊肉新鲜程度，而且传统检测方法操作复杂，效率低，不能满足日常实际需求。高光谱成像技术作为一种快速、无损、高效的检测技术，可以有效地获取冷鲜羊肉腐败过程中表面、内部组成和理化变化信息。提出一种基于改进深度森林算法的冷鲜羊肉新鲜度评价模型，增加特征筛选挖掘与多个评价指标相关的光谱信息，同时增加层增长控制有效防止模型过拟合。采集了0~14天4 ℃贮藏环境中羊肉样本的400~1 000 nm高光谱数据，采用实验室方法测定了样本的挥发性盐基氮（TVB-N）、pH值、菌落总数（TAC）和大肠菌群近似数（ANC）指标值。选择感兴趣区域提取光谱数据，通过S-G平滑滤波法和多元散射校正法对原始光谱数据进行预处理，利用连续投影法提取了18个特征波段。将数据集按照3∶1划分为训练集和测试集；利用本文提出的改进深度森林算法建立新鲜度等级分类模型。结果表明，新鲜度等级分类总体精度为0.985 7，并利用hamming loss、one-error、ranking loss和marco-AUC四种多标记度量指标评价模型性能，分别为0.025 7、0.014 3、0.014 2和0.998 6，均优于传统多标记分类算法也表明，该多指标新鲜度评价模型可用于羊肉新鲜度的快速无损检测，改善了单一新鲜度检测指标模型分类的局限性，为后续高光谱成像技术的多指标无损检测提供了方法。

关键词：高光谱；冷鲜羊肉；新鲜度；多标记分类；深度森林

Abstract：Mutton freshness is affected by many factors, and the detection is generally carried out based on sensory properties, physical and chemical products of decomposition, microbial reproduction and other aspects. However, the freshness detection of mutton based on a single indicator has great limitations and low applicability, and it is not easy to evaluate the mutton freshness comprehensively. Moreover, the traditional detection methods are complex and inefficient, which cannot meet the daily actual needs. As a fast, nondestructive and efficient intelligent detection technology, hyperspectral imaging technology can effectively collect the surface, internal composition and physical and chemical changes in the process of mutton putrefaction. This paper proposes an evaluation model for the freshness of chilled mutton based on the improved deep forest algorithm, which adds feature screening to mine the spectral information related to multiple evaluation indicators. It adds layer growth control to prevent the model from over fitting effectively. This paper collects 400~1 000 nm hyperspectral data of mutton samples stored at 4 ℃ for 0~14 days .Total volatile base nitrogen (TVB-N), pH, total aerobic plate count (TAC) and the approximate number of coliforms (ANC) indicator values are measured by laboratory methods. The representative spectra of mutton samples are extracted in the regions of interest. The original spectral data is preprocessed using the smoothing filtering and multivariate scattering correction methods. 18 spectral feature bands are extracted by using the continuous projection method, and the samples of the training set and the testing set are divided at a ratio of 3∶1. Establishment of the freshness classification model is used in the improved deep forest algorithm proposed in this paper. The results show that the overall accuracy of freshness classification is 0.985 7, and use Hamming loss, One-error, Ranking loss and Marco-AUC multi-label metrics to evaluate the performance of the model, which are 0.025 7, 0.014 3, 0.014 2 and 0.998 6 respectively. Theyare better than the traditional multi-label classification algorithm. The research shows that the multi-indicator freshness classification model can be used for rapid, nondestructive testing of mutton freshness. It improves the limitations of single-indicator model classification and provides a research method for multi-indicator nondestructive testing of subsequent hyperspectral imaging technology.

Key words：Hyperspectraltechnology; Chilledmutton; Freshness; Multi label classification; Deep forest

收稿日期: 2022-07-17 修订日期: 2022-12-17

中图分类号:

TS207.7

基金资助: 国家自然科学基金项目（31960494），内蒙古自治区科技攻关项目（2020GG0169），内蒙古自治区科技重大专项项目（2021ZD0003）资助

通讯作者: 姜新华 E-mail: jiangxh@imau.edu.cn

作者简介: 徐子洋，1998年生，内蒙古农业大学计算机与信息工程学院硕士研究生 e-mail: imau_xzy@163.com

引用本文:

徐子洋，姜新华，白洁，张文婧，李靖. 基于多标记深度森林算法的冷鲜羊肉新鲜度无损检测方法[J]. 光谱学与光谱分析, 2024, 44(02): 580-587.
XU Zi-yang, JIANG Xin-hua, BAI Jie, ZHANG Wen-jing, LI Jing. A Nondestructive Method for Freshness Detection of Chilled Mutton With Multiple Indicators and Improved Deep Forest Algorithm. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2024, 44(02): 580-587.

链接本文:

https://www.gpxygpfx.com/CN/10.3964/j.issn.1000-0593(2024)02-0580-08 或 https://www.gpxygpfx.com/CN/Y2024/V44/I02/580

[1] ZHANG Yu, TIAN Hai-qing, WANG Ke, et al（张珏，田海清，王轲，等）. Journal of China Agricultural University(中国农业大学学报), 2020, 25(5): 100.
[2] Yang Ling, Wu Ting, Liu Yun, et al. Journal of Spectroscopy, 2018, 2018(10): 2314(doi: 10.1055/2018/2413874).
[3] Zhang Xueqi, Jiang Tao, Guo Na, et al. Current Microbiology, 2020, 77: 826.
[4] BAI Xue-bing, MA Dian-kun, ZHANG Meng-jie, et al（白雪冰，马殿坤，张梦杰，等）. Spectroscopy and Spectral Analysis（光谱学与光谱分析）, 2022, 42(7): 1993.
[5] LIU Yan, ZHOU Xin-qi, YU Xiao-feng, et al（刘妍，周新奇，俞晓峰，等）. Journal of Zhejiang University（Agriculture and Life Sciences）[浙江大学学报（农业与生命科学版)], 2020, 46(1): 11.
[6] ZHOU Quan，ZHU Da-zhou，WANG Cheng，et al（周全, 朱大洲, 王成, 等）. Food Science（食品科学）, 2010，31(23): 423.
[7] XIA Yang-yi，ZHAO Luan（夏杨毅，赵鸾）. Journal of Food Science and Biotechnology（食品与生物技术学报），2020，39(12): 6.
[8] Magwaza Lembe S, Opara Umezuruike Linus，Nieuwoudt H，et al. Food & Bioprocess Technology，2012，5(2): 425.
[9] Dellaporta Rick. Journal of Texture Studies, 2008, 39(4): 426.
[10] JIANG Sha，YAN Cai-xia，FAN Xin，et al（姜莎，闫彩霞，范鑫，等）. Meat Industry（肉类工业）, 2021，2: 43.
[11] WANG Hao-yun, SONG Jin, PAN Lei-qing, et al（王浩云，宋进，盘磊庆，等）. Transactions of the Chinese Society of Agricultural Engineering（农业工程学报）, 2020, 36(5): 302.
[12] Jiang Xinhua, Xue Heru, Zhang Lina, et al. Computers and Electronics in Agriculture, 2018, 155(1): 371.
[13] Zhang M L, Zhou Z H. IEEE Transactions on Knowledge & Data Engineering, 2014, 26(8): 1819.
[14] Wu Z H, Qin S W, Nie R C, et al. IEEE Transactions on Neural Networks and Learning Systems, 2022, 33(10): 5200.
[15] Guo Y M, Chung F L, Li G Z, et al. IEEE Access, 2019, 7(4): 103863.
[16] Pereira R B, Plastino A, Zadrozny B, et al. Artificial Intelligence Review, 2018, 49(1): 57.
[17] Zhang M L, Zhou Z H. Pattern Recognition, 2007, 40(7): 2038.
[18] Read J, Pfahringer B, Holmes G, et al. Machine Learning, 2011, 85: 333.
[19] Zhang M L, Zhou Z H. IEEE Transactions on Knowledge & Data Engineering, 2006, 18(10): 1338.
[20] Liu H, Li X, Zhang S. IEEE Transactions on Cybernetics, 2017, 47(2): 499.
[21] XU Jie, LIANG Jiu-zhen, WU Qin, et al（许洁，梁久祯，吴秦，等）. Computer Science（计算机科学）, 2016, 43(1): 35.
[22] Zhou Z H, Feng J. National Science Review，2019, 6(1): 74.
[23] Stojanova D, Ceci M, Appice A, et al. Data Mining and Knowledge Discovery, 2012, 25(2): 378.
[24] Kocev D, Vens C, Struyf J, et al. Pattern Recogn, 2013, 46(3): 817.
[25] National Health and Family Planning Commission of the People's Republic of China(国家卫生和计划生育委员会). GB/5009.228—2016 National Food Safety Standard Determination of Total Volatile Base Nitrogen in Food(GB/5009.228—2016 食品国家安全标准食品中挥发性盐基氮的测定), 2016.
[26] National Health and Family Planning Commission of the People's Republic of China(国家卫生和计划生育委员会). GB/5009.237—2016 National Food Safety Standard Determination of pH of Food(GB/5009.237—2016 食品国家安全标准食品pH值的测定), 2016.
[27] National Health and Family Planning Commission of the People's Republic of China(国家卫生和计划生育委员会). GB/4789.2—2022 National Food Safety Standard Food Microbiological Examination: Aerobic Plate Count(GB/4789.2—2022 食品国家安全标准食品微生物学检验菌落总数测定), 2022.
[28] National Health and Family Planning Commission of the People's Republic of China(国家卫生和计划生育委员会). GB/4789.3—2016 National Food Safety Standard Food Microbiological Examination: Approximate Number of Coliforms(GB/4789.3—2016 食品国家安全标准食品微生物学检验大肠杆菌计数), 2016.