• Laser & Optoelectronics Progress
  • Vol. 55, Issue 6, 063002 (2018)
Zhonghan Zhou1、2、1; 2; , Xueyong Tian3、3; , Lanxiang Sun1、1; , Peng Zhang1、2、1; 2; , Zhiwei Guo1、4、1; 4; , and Lifeng Qi1、1;
Author Affiliations
  • 1 Shenyang Institute Automation, Chinese Academy of Sciences, Shenyang, Liaoning 110016, China
  • 2 University of Chinese Academy of Sciences, Beijing 100049, China
  • 3 Central Research Institute, SIASUN Robot and Automation Co., Ltd., Shenyang, Liaoning 110168, China
  • 4 College of Information Science and Engineering, Northeastern University, Shenyang, Liaoning 110819, China
  • show less
    DOI: 10.3788/LOP55.063002 Cite this Article Set citation alerts
    Zhonghan Zhou, Xueyong Tian, Lanxiang Sun, Peng Zhang, Zhiwei Guo, Lifeng Qi. Identification of Aluminum Alloy Grades by Fiber-Laser Induced Breakdown Spectroscopy Combined with Support Vector Machine[J]. Laser & Optoelectronics Progress, 2018, 55(6): 063002 Copy Citation Text show less
    Fiber-LIBS experimental system
    Fig. 1. Fiber-LIBS experimental system
    Schematic of control sequence of laser and spectrometer
    Fig. 2. Schematic of control sequence of laser and spectrometer
    Aluminum alloy samples
    Fig. 3. Aluminum alloy samples
    Prediction results of original spectral test sets
    Fig. 4. Prediction results of original spectral test sets
    Invalid spectra, saturated spectra and normal spectra
    Fig. 5. Invalid spectra, saturated spectra and normal spectra
    Prediction results of filtered spectral test sets
    Fig. 6. Prediction results of filtered spectral test sets
    Prediction results of normalized spectral test sets
    Fig. 7. Prediction results of normalized spectral test sets
    Prediction results of spectral test sets after filter, normalization and principal component analysis
    Fig. 8. Prediction results of spectral test sets after filter, normalization and principal component analysis
    Verification of SVM with different inputs using leave-one-out. (a) Prediction accuracy; (b) modeling time
    Fig. 9. Verification of SVM with different inputs using leave-one-out. (a) Prediction accuracy; (b) modeling time
    CategoryGradeSerial number
    1BBA.AlSi7MgCu(0.5)Fe(0.2)1, 2
    2DC.360Y.63, 4, 5, 6, 7
    3HB.SF368, 9, 10,11, 12, 13
    4WD.AlSi7MgCu14, 15, 16
    5WD.AlSi9Cu3(Fe)17, 18, 19
    6YD.ZAlSi7Mg0.320, 21, 22, 23, 24
    Table 1. Grades and serial number of aluminum alloy samples
    InputAverage prediction accuracy /%Mean modeling time /s
    Original spectra92.3447.91
    Filter98.698.06
    Filter+normalization99.797.45
    Filter+normalization+PCA99.830.14
    Table 2. Average prediction accuracy and mean modeling time of SVM with different inputs
    Zhonghan Zhou, Xueyong Tian, Lanxiang Sun, Peng Zhang, Zhiwei Guo, Lifeng Qi. Identification of Aluminum Alloy Grades by Fiber-Laser Induced Breakdown Spectroscopy Combined with Support Vector Machine[J]. Laser & Optoelectronics Progress, 2018, 55(6): 063002
    Download Citation