Visual Inspection of Food Packaging Paper by X-Ray Fluorescence Spectroscopy Combined with Deep Learning Algorithm

Qi Guo; Hong Jiang; Jinjie Yang; Kenan Wu; Ji Man

doi:10.3788/LOP202259.0434001

Journals >Laser & Optoelectronics Progress >Volume 59 >Issue 4 >Page 0434001 > Article

Laser & Optoelectronics Progress
Vol. 59, Issue 4, 0434001 (2022)

Visual Inspection of Food Packaging Paper by X-Ray Fluorescence Spectroscopy Combined with Deep Learning Algorithm

Qi Guo¹, Hong Jiang^1、*, Jinjie Yang¹, Kenan Wu², and Ji Man³

Author Affiliations

¹Institute of Criminal Investigation, People's Public Security University of China, Beijing 100038, China

²Institute of Computer Science and Technology, Wuhan University of Technology, Wuhan , Hubei 430070, China

³Beijing Huayi Honrizon Technology Co., Ltd., Beijing 100123, China

show less

DOI: 10.3788/LOP202259.0434001 Cite this Article Set citation alerts

Qi Guo, Hong Jiang, Jinjie Yang, Kenan Wu, Ji Man. Visual Inspection of Food Packaging Paper by X-Ray Fluorescence Spectroscopy Combined with Deep Learning Algorithm[J]. Laser & Optoelectronics Progress, 2022, 59(4): 0434001 Copy Citation Text

show less

Fig. 1. Manual classification results

Download full size

Fig. 2. Tree diagram of clustering results

Download full size

Fig. 3. Line chart of polymerization coefficient

Download full size

Fig. 4. Comparison of 2D renderings of dimension reduction algorithms. (a) PCA; (b) tSNE

Download full size

Fig. 5. Comparison of 3D renderings of dimension reduction algorithms. (a) PCA; (b) tSNE

Download full size

Sample No.	Ca	Fe	Sn	Zn	Cl	Ti	Sample No.	Ca	Fe	Sn	Zn	Cl	Ti
1^#	1341	336	265	25	972	4315	23^#	17091	528	123	0	3046	0
2^#	265	185	152	6	4267	0	24^#	430	215	137	14	2037	3568
3^#	710	280	125	16	642	25	25^#	144224	5997	286	156	0	957
4^#	22620	352	122	14	2230	0	26^#	36102	950	290	48	2467	0
5^#	10312	315	75	13	4767	0	27^#	57704	1754	288	53	0	142
6^#	338277	7237	243	152	27057	7386	28^#	86662	1987	303	52	3740	145
7^#	974	593	223	15	1610	0	29^#	149657	3272	374	67	5898	247
8^#	2418	641	191	16	629	0	30^#	62564	1139	227	28	754	128
9^#	9848	401	316	17	8436	472	31^#	81827	1618	430	37	10852	116
10^#	9252	313	187	13	2509	988	32^#	4417	300	288	16	20487	0
11^#	50212	1358	176	47	0	0	33^#	1019	417	212	18	2898	28
12^#	90821	4351	253	89	0	234	34^#	653951	7182	546	1378	9948	821
13^#	480	224	199	13	885	0	35^#	519	534	93	0	2329	0
14^#	48275	1207	220	30	1688	0	36^#	348362	1834	261	57	8147	2457
15^#	168161	1763	543	60	2121	0	37^#	278	231	150	28	5371	0
16^#	182537	5715	355	136	2023	249	38^#	374	190	111	0	3846	0
17^#	58463	2015	266	59	1798	155	39^#	87355	988	522	56	4459	0
18^#	110658	3077	271	95	0	216	40^#	130055	1084	151	33	1650	0
19^#	284881	7290	471	259	3050	610	41^#	99722	1163	296	37	1839	0
20^#	206208	5276	435	114	2608	772	42^#	102456	816	165	30	4822	0
21^#	127196	3536	309	99	0	306	43^#	46252	642	125	15	1682	0
22^#	87681	2542	281	105	0	319	44^#	29072	1869	75	11	934	4642

Table 1. X-ray fluorescence detection results

Dimension	PCA	tSNE
2D	0.25527	0.336786
3D	0.14772	0.0917815

Table 2. Comparison of data loss values of two algorithms

Dataset	Category	Number of output result samples of each category					Accuracy /%
Dataset	Category	4	5	6	7	9	Accuracy /%
Train	4	3	0	0	0	0	100.0
	5	0	3	0	0	0	100.0
	6	0	0	24	0	0	100.0
	7	0	0	0	2	0	100.0
	9	0	0	0	0	3	100.0
	Overall percentage /%	8.6	8.6	68.6	5.7	8.6	100.0
Test	4	1	0	0	0	0	100.0
	5	0	1	0	0	0	100.0
	6	0	0	5	0	0	100.0
	7	0	0	0	1	0	100.0
	9	0	0	0	1	0	0.0
	Overall percentage /%	11.1	11.1	55.6	22.2	0.0	88.9

Table 3. Artificial neural network outputs classification results

Download Citation

Set citation alerts for the article

Tools

Set citation alerts for the article

Save the article for my favorites

Paper Information