Collaborative Representation-Based Binary Hypothesis Model with Multi-features Learning for Target Detection in Hyperspe
- PDF / 3,702,225 Bytes
- 16 Pages / 595.276 x 790.866 pts Page_size
- 95 Downloads / 175 Views
RESEARCH ARTICLE
Collaborative Representation-Based Binary Hypothesis Model with Multi-features Learning for Target Detection in Hyperspectral Imagery Chunhui Zhao1 • Wei Li1,2 Received: 30 January 2016 / Accepted: 31 January 2018 / Published online: 22 May 2018 Ó Indian Society of Remote Sensing 2018
Abstract In this paper, we propose a collaborative representation-based binary hypothesis model with multi-features learning (CRTDBH-MTL) for target detection in hyperspectral imagery. The proposed method contained the following aspects. First, two complementary features extracted by different algorithms are implemented for describing hyperspectral imageries. Next, we apply these features into the unified collaborative representation-based binary hypothesis model (CRTDBH) to acquire a collaborative vector (CV) for each feature. Once the CV is obtained, the sample can be sparsely represented by the training samples from the background-only dictionary under the null hypothesis and the training samples from the target and background dictionaries under the alternative hypothesis. Finally, spatial correlation and spectral similarity of adjacent neighboring pixels are exploited to improve the detection performance. The experimental results suggest that the proposed algorithm shows an outstanding detection performance. Keywords Hyperspectral imagery (HSI) Collaborative representation Binary hypothesis Multi-features learning Target detection
Introduction Hyperspectral imageries (HSIs) spanning the visible to the infrared spectrum with a wealth spectral information, afford excellent spectral differences between kinds of materials, therefore supporting improved target detection accuracy (Zhao et al. 2016; Fauvel et al. 2008). In the hyperspectral target detection method, the class label of each sample, denoted by a vector whose entries correspond to responses to spectral bands, is determined by a given training set from each class (Licciardi and Chanussot 2015; Chen et al. 2011b). Support vector machines (SVMs) (Lin et al. 2007; Fauvel et al. 2008) which can solve supervised binary detection problems are widely used in the & Chunhui Zhao [email protected] 1
College of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, China
2
Department of Earth and Atmospheric Sciences, Alberta Centre for Earth Observation Sciences, University of Alberta, Edmonton T6G 2E3, Canada
hyperspectral image processing. Sparse representation target detection (SRTD) method has been widely used for hyperspectral target detection (Chen et al. 2011b). In SRTD, the test pixel can be described by some training atoms from the over-completed dictionary (Li and Seshia 2013). Nowadays, the collaborative representation method (CR) Zhang et al. (2012a), which describes pixels by a linear combination of some training samples, acquired good results in hyperspectral target detection (Chen et al. 2011a). However, the CR with single feature is not good at distinguishing objects since one single feature could onl
Data Loading...
