Share Email Print

Proceedings Paper • new

A DBN based anomaly targets detector for HSI
Author(s): Ning Ma; Shaojun Wang; Jinxiang Yu; Yu Peng
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Due to the assumption that Hyperspectral image (HSI) should conform to Gaussian distribution, traditional Mahalanobis distance-based anomaly targets detectors perform poor because the assumption may not always hold. In order to solve those problems, a deep learning based detector, Deep Belief Network(DBN) anomaly detector(DBN-AD), was proposed to fit the unknown distribution of HSI by energy modeling, the reconstruction errors of this encode-decode processing are used for discriminating the anomaly targets. Experiments are implemented on real and synthesized HSI dataset which collection by Airborne Visible Infra-Red Imaging Spectrometer (AVIRIS). Comparing to classic anomaly detector, the proposed method shows better performance, it performs about 0.17 higher in Area Under ROC Curve (AUC) than that of Reed-Xiaoli detector(RXD) and Kernel-RXD (K-RXD).

Paper Details

Date Published: 24 October 2017
PDF: 6 pages
Proc. SPIE 10458, AOPC 2017: 3D Measurement Technology for Intelligent Manufacturing, 104581Z (24 October 2017); doi: 10.1117/12.2285766
Show Author Affiliations
Ning Ma, Harbin Institute of Technology (China)
Shaojun Wang, Harbin Institute of Technology (China)
Jinxiang Yu, Harbin Institute of Technology (China)
Yu Peng, Harbin Institute of Technology (China)

Published in SPIE Proceedings Vol. 10458:
AOPC 2017: 3D Measurement Technology for Intelligent Manufacturing
Wolfgang Osten; Anand Krishna Asundi; Huijie Zhao, Editor(s)

© SPIE. Terms of Use
Back to Top