Quantitative Detection of Mutton Hardness Based on Twice Iterative Monte Carlo Method

Xue-bing BAI; Xin-xing LI; Xiao-shuan ZHANG; Hai-ling LUO; Ze-tian FU

doi:10.3964/j.issn.1000-0593(2021)07-2057-07

Journals >Spectroscopy and Spectral Analysis >Volume 41 >Issue 7 >Page 2057 > Article

Spectroscopy and Spectral Analysis
Vol. 41, Issue 7, 2057 (2021)

Quantitative Detection of Mutton Hardness Based on Twice Iterative Monte Carlo Method

Xue-bing BAI^1、*, Xin-xing LI^1、1;, Xiao-shuan ZHANG^2、2;, Hai-ling LUO^3、3;, and Ze-tian FU^{2、2; *;}

Author Affiliations

¹1. Beijing Laboratory of Food Quality and Safety, College of Information and Electrical Engineering, China Agricultural University, Beijing 100083, China

²2. College of Engineering, China Agricultural University, Beijing 100083, China

³3. State Key Laboratory of Animal Nutrition, College of Animal Science and Technology, China Agricultural University, Beijing 100083, China

show less

DOI: 10.3964/j.issn.1000-0593(2021)07-2057-07 Cite this Article

Xue-bing BAI, Xin-xing LI, Xiao-shuan ZHANG, Hai-ling LUO, Ze-tian FU. Quantitative Detection of Mutton Hardness Based on Twice Iterative Monte Carlo Method[J]. Spectroscopy and Spectral Analysis, 2021, 41(7): 2057 Copy Citation Text

show less

Fig. 1. The development trends of muttons' hardness under different storage temperatures

Download full size | View in the Article

Fig. 2. Original spectrum of mutton sample
(a): 400~1 020 nm; (b): 860~1 700 nm

Download full size | View in the Article

Fig. 3. Mutton hardness outliers' detection result by 2 iteration MC sampling
(a): 400~950 nm 1st MC; (b): 900~1 650 nm 1st MC; (c): 400~950 nm 2nd MC; (d): 900~1 650 nm 2nd MC

Download full size | View in the Article

Fig. 4. Optimal wavelengths of mutton's hardness by RC
(a): 400~900 nm; (b): 900~1 650 nm

Download full size | View in the Article

Fig. 5. Mutton's hardness prediction results of PLSR
(a): 400~950 nm prediction results; (b): 900~1 650 nm prediction results

Download full size | View in the Article

		400~950 nm			900~1 650 nm
		样本编号	$R_{CV}^{2}$	RMSE_CV	样本编号	$R_{CV}^{2}$	RMSE_CV
1^stMC	异常样本	2, 35, 74	0.915 9	61.657 0	140	0.957 9	40.613 1
1^stMC	可疑样本	3, 4, 83, 109	-	-	4, 5, 12, 138	-	-
2^nd MC		3, 4, 104	0.936 2	53.697 4	4, 12	0.964 7	37.191 3
样本二次检测		剔除83	0.940 0	52.090 5	剔除138	0.976 9	37.442 8
样本二次检测		剔除109	0.933 7	54.739 3	剔除5	0.982 8	37.040 3

Table 1. Detection results of mutton hardness outliers based on 2 iteration MC sampling

View in the Article

预处理	400~950 nm				900~1 650 nm
预处理	$R_{C}^{2}$	RMSE_C	$R_{P}^{2}$	RMSE_P	$R_{C}^{2}$	RMSE_C	$R_{P}^{2}$	RMSE_P
原始光谱	0.945 7	49.565 8	0.934 1	55.716 2	0.966 8	35.784 5	0.964 6	37.590 1
S-G平滑	0.945 6	49.587 0	0.934 5	55.495 1	0.966 0	36.179 6	0.962 7	34.612 1
二阶求导	0.947 4	48.758 8	0.933 0	56.088 7	0.970 2	34.105 4	0.968 5	31.803 2
MSC	0.959 3	41.595 1	0.947 2	47.789 9	0.979 2	29.660 7	0.978 3	30.590 1
SNV	0.951 9	46.619 1	0.932 6	56.387 2	0.975 7	31.389 2	0.972 2	30.716 7

Table 2. Results of spectral data preprocessing

View in the Article

	模型	LVs	校正集		交叉验证集		预测集
	模型	LVs	$R_{C}^{2}$	RMSE_C	$R_{CV}^{2}$	RMSE_CV	$R_{P}^{2}$	RMSE_P	RPD
400~950 nm	全波长	8	0.959 3	41.595 1	0.950 0	46.357 3	0.947 2	47.789 9	4.350
400~950 nm	RC	7	0.966 0	38.039 7	0.958 2	42.780 8	0.951 9	46.602 7	4.425
900~1 650 nm	全波长	8	0.979 2	29.660 7	0.978 5	29.855 3	0.978 3	30.590 1	6.717
900~1 650 nm	RC	6	0.987 3	22.418 6	0.986 6	23.480 9	0.985 0	24.397 0	7.524

Table 3. Mutton hardness prediction results of PLSR models

Download Citation

Save the article for my favorites

Paper Information