Effective Wavelengths Selection of Hyperspectral Images of Plastic Films in Cotton
This research was conducted to investigate the application of detecting plastic films in cotton using visible and near-infrared hyperspectral imaging. A line-scan hyperspectral imaging system (326–1100 nm) was used to detect plastic films mixed with cotto
- PDF / 2,283,908 Bytes
- 9 Pages / 439.37 x 666.142 pts Page_size
- 39 Downloads / 158 Views
3
1 College of Information and Electrical Engineering, China Agricultural University, Beijing 100083, China [email protected] 2 College of Computer and Information Engineering, Tianjin Agricultural University, Tianjin 300384, China Key Laboratory of Agricultural Information Acquisition Technology, Ministry of Agriculture, Beijing 100083, China
Abstract. This research was conducted to investigate the application of detecting plastic films in cotton using visible and near-infrared hyperspectral imaging. A line-scan hyperspectral imaging system (326–1100 nm) was used to detect plastic films mixed with cotton which was an important quality issue. Hyperspectral reflectance images were acquired and difference spectra of cotton and plastic films were extracted and analyzed to determine the dominant wavelengths. Also, as one of the most commonly used methods for dimensionality reduction, principal component analysis (PCA) was chosen to process the hyperspectral images. Afterwards, effective wavelengths were selected by analyzing the first three principal components (PCs) and six single-band images at 473.24 nm, 497.29 nm, 530.6 nm, 670.81 nm, 674.71 nm, and 955.68 nm were extracted respectively. Finally, the selected wavelengths were validated to prove the effectiveness. The results indicated that the selected wavelengths could be able to detect plastic films in cotton instead of the whole wavelengths. Keywords: Hyperspectral imaging Difference spectra Principal component analysis Wavelengths selection
1 Introduction The quality of cotton products is easily and seriously affected by foreign matter mixed into cotton during spinning, weaving, and dyeing. As one kind of foreign matter, plastic films are widely used to preserve soil temperature and moisture in China when growing cotton and make it a problem that a great many plastic films might be mixed with cotton when harvesting. They are easily to be broken into lots of very small pieces during spinning [1], which make removal of them difficult and make breakability of cotton yarn increase. Therefore, it is crucial for us to detect and remove plastic films in cotton rapidly and accurately.
© IFIP International Federation for Information Processing 2016 Published by Springer International Publishing AG 2016. All Rights Reserved D. Li and Z. Li (Eds.): CCTA 2015, Part I, IFIP AICT 478, pp. 519–527, 2016. DOI: 10.1007/978-3-319-48357-3_48
520
H. Zhang et al.
For the past two decades, computer vision technique has been applied to identify bark in cotton and determine the gravimetric bark content in cotton [2], conduct automated visual inspection of cotton quality [3], measure interlace of intermingled and false-twist textured yarns [4], detect structural defects in textiles [5], inspect and classify different types of foreign fibers [6, 7], etc. Furthermore, research has also been conducted on plastic films. Fang et al. [1] used an online visual detection machine to acquire images of plastic films and proposed a new method to identify plastic films. The result showed that the
Data Loading...
