Noise Robust Hyperspectral Image Classification With MNF-Based Edge Preserving Features




Edge preserving features (EPFs), hyperspectral image (HSI) classification, minimum noise fraction (MNF), principal component analysis (PCA), support vector machine (SVM)


Hyperspectral image (HSI) classification is an important topic in remote sensing. In this paper, we improve the principal component analysis (PCA)-based edge preserving features (EPFs) for HSI classification. We select to use minimum noise fraction (MNF) instead of PCA to reduce the dimensionality of the hyperspectral data cube to be classified. We keep all the rest steps from the PCA-based EPFs for HSI classification. Since MNF can preserve fine features of a HSI data cube better than PCA, our new method can outperform PCA-EPFs for HSI classification significantly. Experimental results show that our new method performs better than the PCA-based EPFs under such noisy environment as Gaussian white noise and shot noise. In addition, our MNF+EPFs outperform the PCA+EPFs even when no noise is added to the HSI data cubes for most testing cases, which is very desirable in remote sensing.

Author Biographies

Guangyi Chen, Concordia University

Guang Yi Chen holds a B.Sc. in Applied Mathematics, an M.Sc. in Computing Mathematics, an M.Sc. in Computer Science, and a Ph.D. in Computer Science. During his graduate and postdoctoral studies in Canada, he was awarded many prestigious fellowships. He has published over sixty-five scientific journal papers in his fields and holds two granted USA patents in image processing. He is currently affiliated to the Department of Computer Science and Software Engineering, Concordia University, Montreal, Quebec, Canada. He is the world's top 2% scientist ranked by Stanford University. His research interests include pattern recognition, image processing, machine learning, artificial intelligence, remote sensing, and scientific computing.

Adam Krzyzak

Adam Krzyzak received the M.Sc. and Ph.D. degrees in computer engineering from the Wrocław University of Science and Technology, Poland, in 1977 and 1980, respectively, and D.Sc. degree (habilitation) in computer engineering from the Warsaw University of Technology, Poland in 1998. In 2003 he received the Title of Professor from the President of the Republic of Poland. Since 1983, he has been with the Department of Computer Science and Software Engineering, Concordia University, Montreal, Canada, where he is currently a full professor. He published over 350 papers on neural networks, pattern recognition, nonparametric estimation, image processing, computer vision and control. He has been an associate editor of IEEE Transactions on Neural Networks and IEEE Transactions on Information Theory and is presently an Associate Editor-in-Chief of the Pattern Recognition Journal. He was co-editor of the book Computer Vision and Pattern Recognition (Singapore: World Scientific, 1989) and is a co-author of the book A Distribution-Free Theory of Nonparametric Regression, New York: Springer, 2002. He is a Fellow of the IEEE and a Fellow of IAPR.

Shen-en Qian, Canadian Space Agency

Shen-En Qian received his Ph.D. degrees in telecommunication and electronic systems in 1990. He is a principal scientist and technical lead of space missions at the Canadian Space Agency since 1995. He is the sole author of four books on optical satellites, system design and their signal processing published in USA and UK. He is a co-author of four other books. He holds 35 granted patents worldwide developed in Canadian government laboratory. He published over 120 papers in the areas of optical spacecraft payloads, space technologies for satellite missions and deep space exploration, remote sensing, satellite signal processing and enhancement, satellite data compression, data handling and establishment of international standards for spacecraft data systems. He is a fellow of the International Society of Optics and Photonics (SPIE), a fellow of Canadian Academy of Engineering (CAE). He is an Associate Editor of SPIE’s Journal of Applied Remote Sensing, and an Associate Editor of IEEE Journal on Miniaturization for Air and Space Systems (J-MASS). He is an adjunct professor at York University. He received IEEE Canada Outstanding Engineer Award in 2019. The Governor General of Canada awarded him the Public Service Award of Excellence in 2016 for his exceptional contributions in advancing Canadian space programs. He received the Award of Canadian Government Invention for his multiple patents for space missions in 2004. He received the Marie Curie Award from European Community in 1992.



Camps-Valls G and Bruzzone L (2005), Kernel-based methods for hyperspectral image classification, IEEE T GEOSCI REMOTE, 43:1351-62.

Chen GY, Bui TD, Quach KG and Qian SE (2014), Denoising hyperspectral imagery using principal component analysis and block matching 4D filtering, CAN J REMOTE SENS, 40:60-7.

Chen GY and Qian SE (2011), Denoising of hyperspectral imagery using principal component analysis and wavelet shrinkage, IEEE T GEOSCI REMOTE, 49:973-80.

Chen Y, Lin Z, Zhao X, Wang G and Gu Y (2014), Deep learning-based classification of hyperspectral data, IEEE J-STARS, 7:2094-107.

Chen Y, Nasrabadi NM and Tran TD (2011), Hyperspectral image classification using dictionary-based sparse representation, IEEE T GEOSCI REMOTE, 49:3973-85.

Chen Y, Nasrabadi NM and Tran TD (2013), Hyperspectral image classification via kernel sparse representation, IEEE T GEOSCI REMOTE, 51:217-31.

Cheng G, Zhu, F, Xiang S, Wang Y and Pan X (2016), Semisupervised hyperspectral image classification via discriminant analysis and robust regression, IEEE J-STARS, 9:595-608.

Cortes C and Vapnik VN (1995), Support-vector networks, Machine Learning, 20:273–97.

Green AA, Berman M, Switzer P and Craig MD (1988), A transformation for ordering multispectral data in terms of image quality with implications for noise removal, IEEE T GEOSCI REMOTE, 26:65-74.

Jolliffe IT (2002), Principal Component Analysis, second edition (Springer).

Kang XD, Li S and Benediktsson JA (2014), Spectral-spatial hyperspectral image classification with edge-preserving filtering, IEEE T GEOSCI REMOTE, 52:2666-77.

Kang X, Xiang X, Li S and Benediktsson JA (2017), PCA-Based Edge-Preserving Features for Hyperspectral Image Classification, IEEE T GEOSCI REMOTE, 55:7140-51.

M. Fauvel, J. Chanussot and J. A. Benediktsson, “A spatial–spectral kernel-based approach for the classification of remote-sensing images,” PATTERN RECOGN, vol. 45, no. 1, pp. 381-392, 2012.

Li J, Bioucas-Dias JM and Plaza A (2013), Spectral-spatial classification of hyperspectral data using loopy belief propagation and active learning, IEEE T GEOSCI REMOTE, 51:844-56.

Li J, Marpu PR, Plaza A, Bioucas-Dias JM and Benediktsson JA (2013), Generalized composite kernel framework for hyperspectral image classification, IEEE T GEOSCI REMOTE, 51:4816-29.

Liu H, Xia K, Li T, Ma J and Owoola E (2020). Dimensionality reduction of hyperspectral images based on improved spatial–spectral weight manifold embedding, Sensors 20:4413.

Luo GC, Chen GY, Tian L, Qin K and Qian SE (2016), Minimum noise fraction versus principal component analysis as a preprocessing step for hyperspectral imagery denoising, CAN J REMOTE SENS, 42:106-16.

Melgani F and Bruzzone I (2004), Classification of hyperspectral remote sensing images with support vector machines, IEEE T GEOSCI REMOTE, 42:1778-90.

Zhou F, Hang R, Liu Q and Yuan X (2019), Hyperspectral image classification using spectral-spatial LSTMs, Neurocomputing, 328:39-47.




How to Cite

Chen, G., Krzyzak, A., & Qian, S.- en. (2023). Noise Robust Hyperspectral Image Classification With MNF-Based Edge Preserving Features. Image Analysis and Stereology, 42(2), 93–99.



Original Research Paper