|
| Optimization of Pretreatment Methods for Plastic Sorting Technology Based on Near-infrared Spectroscopy |
| |
| View Full Text Download reader |
| DOI： |
| KeyWord:plastic sorting near-infrared spectroscopy（NIR） preprocessing combination predictive analysis |
| Author | Institution |
| LI Jia-shuai，XUE Lian-lian，WANG Kai，LIU Jun-cheng，WU Han，LI Hua-qing，YIN Feng-fu |
1. School of Mechanical and Electrical Engineering，Qingdao University of Science and Technology， Qingdao ，China; 2. School of Information Engineering，Qingdao Vocational College of Engineering，Qingdao ，China |
|
| Hits: 604 |
| Download times: 1140 |
| Abstract: |
| In order to identify different types of plastics，the near-infrared spectral data of four types of plastics，i.e．nylon（PA），polypropylene（PP），polystyrene（PS） and polyvinyl chloride（PVC），were collected．Meanwhile，according to the existing noise，baseline and optical path issues during spectral data acquisition，pretreatment combination optimization methods were investigated，based on 3-point Savitzky－Golay convolution smoothing（S－G），first derivative（FD），second derivative（SD），standard normal variable transformation（SNV） and multivariate scattering correction（MSC）．Furthermore，the characteristic wavelength was extracted by competitive adaptive reweighting sampling（CARS），and a model was established using support vector machine（SVM）．The results showed that among all preprocessing methods，the preprocessing combination S－G + FD + SNV obtains the best results，while the average accuracy of the S－G + FD + SNV + SVM model is as high as 96.67%，and the average accuracies of its training set and validation set are both 100%．The mentioned-above pretreatment combination optimization method could provide a reference for the identification on four common plastics. |
| Close |