Skip to Main content Skip to Navigation
Conference papers

Study on Quick Identify of the Brand of Seabuckthorn Juice Based on PCA and SVM

Abstract : To achieve the Seabuckthorn juice brand non-destructively, we put forward a fast identify method based on Visible_near infrared reflectance (NIR) spectroscopy. We use the Field Spec 3 spectroradiometer to collect 40 sample spectra data of the three kinds of Seabuckthorn juice separately. the sample data was preprocessed by Using of average smoothing method and multiplicative scatter correction (MSC) method. Then principal component analysis (PCA) was used to process the spectral data after pretreatment. The samples were divided into 90 model samples and 30 prediction samples, the sample of eight modeling data as input variables of the support vector machine (SVM) to build SVM model, and to identification juice brands. 30 unknown samples of the three brands were predicted for classification, and the results showed that the SVM model on the identification seabuckthorn juice brand has achieved a 99.9% accuracy. Therefore, near infrared spectroscopy coupled with principal component analysis and SVM can be quickly and accurately distinguish the brand of seabuckthorn juice.
Document type :
Conference papers
Complete list of metadata

Cited literature [16 references]  Display  Hide  Download

https://hal.inria.fr/hal-01361002
Contributor : Hal Ifip <>
Submitted on : Tuesday, September 6, 2016 - 3:13:44 PM
Last modification on : Tuesday, September 6, 2016 - 4:06:06 PM
Long-term archiving on: : Wednesday, December 7, 2016 - 2:18:44 PM

File

978-3-642-27278-3_37_Chapter.p...
Files produced by the author(s)

Licence


Distributed under a Creative Commons Attribution 4.0 International License

Identifiers

Citation

Zhipeng Liu, Shujuan Zhang. Study on Quick Identify of the Brand of Seabuckthorn Juice Based on PCA and SVM. 5th Computer and Computing Technologies in Agriculture (CCTA), Oct 2011, Beijing, China. pp.351-358, ⟨10.1007/978-3-642-27278-3_37⟩. ⟨hal-01361002⟩

Share

Metrics

Record views

69

Files downloads

230