Wheat Quality Recognition Based on Watershed Algorithm and Kernel Partial Least Squares

Author

An, Ning ; Cong, Pei-sheng ; Zhu, Zhong-liang

Author_Institution

Dept. of Chem., Tongji Univ., Shanghai, China

Volume

2

fYear

2009

fDate

14-16 Aug. 2009

Firstpage

265

Lastpage

269

Abstract

Wheat quality recognition is depended on its shape and color characteristics. Watershed algorithm often can be used to extract complete particles images from the wheat photos, and get their important characteristics. In this paper, Kernel PLS (KPLS) algorithm was used to build a model for wheat kernel classification. A 3-layer back propagation artificial neural network (ANN) was also used for the same data set. The results showed that feature extraction techniques based on high performance watershed algorithm was reliable and high-speed. Average classification accuracy of KPLS and ANN for test set reached 98.00% and 97.00%.

Keywords

agriculture; backpropagation; feature extraction; image classification; quality control; Kernel PLS algorithm; Kernel partial least squares; back propagation artificial neural network; complete particles image extraction; feature extraction; watershed algorithm; wheat kernel classification; wheat quality recognition; Artificial neural networks; Chemistry; Digital cameras; Image segmentation; Kernel; Least squares methods; Pattern recognition; Principal component analysis; Testing; Training data; Kernel-PLS (KPLS); image processing; quality recognition; wheat kernels;

fLanguage

English

Publisher

ieee

Conference_Titel

Natural Computation, 2009. ICNC '09. Fifth International Conference on

Conference_Location

Tianjin

Print_ISBN

978-0-7695-3736-8

Type

conf

DOI

10.1109/ICNC.2009.126

Filename

5365376