Compression and noise reduction of hyperspectral images using non-negative tensor decomposition and compressed sensing

Hassanzadeh, Shirin; Karami, Alimohammad

doi:10.5721/eujrs20164931

Cited by 12 publications

(11 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Sorber et al introduced a quasi-Newton optimization algorithm that iteratively improves the initial guess by using a quasi-Newton method from the nonlinear cost function [17]. Hassanzadeh and karami proposed a block coordinate descent search based algorithm [18], which updates the factor matrices initialized by using compressed sensing. Instead of employing SVD for unfolding matrices, Phan et al proposed a fast algorithm based on a Crank-Nicholson-like technique, which has a lower computational cost in a single step compared with HOOI [19].…”

Section: Related Workmentioning

confidence: 99%

“…More information about Urban dataset is provided in [21]. To compare the performances of the proposed algorithm with previous algorithms, we evaluated the relative errors and the execution times with the algorithm from HOOI, a Crank-Nicholson-like algorithm for HOSVD (CrNc henceforce) [19]; a quasi-Newton-based nonlinear least squares algorithm (henceforce HOSVD_NLS) [17]; and a method based on block coordinate descent search [18] which is a slight modification of the algorithm described in [23] (henceforth BCD-CD). Here, the relative errors, denoted as relerr, are defined such that…”

mentioning

confidence: 99%

See 1 more Smart Citation

An Efficient Compressive Hyperspectral Imaging Algorithm Based on Sequential Computations of Alternating Least Squares

Lee

2019

Remote Sensing

View full text Add to dashboard Cite

Hyperspectral imaging is widely used to many applications as it includes both spatial and spectral distributions of a target scene. However, a compression, or a low multilinear rank approximation of hyperspectral imaging data, is required owing to the difficult manipulation of the massive amount of data. In this paper, we propose an efficient algorithm for higher order singular value decomposition that enables the decomposition of a tensor into a compressed tensor multiplied by orthogonal factor matrices. Specifically, we sequentially compute low rank factor matrices from the Tucker-1 model optimization problems via an alternating least squares approach. Experiments with real world hyperspectral imaging revealed that the proposed algorithm could compute the compressed tensor with a higher computational speed, but with no significant difference in accuracy of compression compared to the other tensor decomposition-based compression algorithms.

show abstract

Section: Related Workmentioning

confidence: 99%

mentioning

confidence: 99%

An Efficient Compressive Hyperspectral Imaging Algorithm Based on Sequential Computations of Alternating Least Squares

Lee

2019

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Bu durumda, sıkıştırma sonrasında veri kalitesinin düşmemesinin yanı sıra uydu üzerinde çok güçlü donanım ve kaynaklar olmadığından, yöntemin düşük karmaşıklığa sahip olması da beklenmektedir. Literatürdeki çalışmalarda kullanılan sıkıştırma yapılarının yüksek işlemsel karmaşıklık içermesinden dolayı işlem sürelerinin oldukça uzun olduğu belirtilmektedir [22]- [24]. Örneğin, [22]'de yapılan çalışmada PCA dönüşümü ile DCT işlemsel karmaşıklık açısından karşılaştırılmıştır.…”

Section: Introductionunclassified

“…Örneğin, [22]'de yapılan çalışmada PCA dönüşümü ile DCT işlemsel karmaşıklık açısından karşılaştırılmıştır. Karşılaştırmada, PCA'in hesabında kovaryans matrisi ve özdeğer belirleme adımlarından dolayı hesapsal karmaşıklığın DCT'ye göre oldukça büyük olduğu belirtilmiştir [22]. Ayrıca, PCA dönüşümü veri bağımlı olduğundan her veri için ayrıca hesaplanması gerekmektedir [22].…”

Section: Introductionunclassified

Compression of hyperspectral images using automatic adaptive luminance transform and 3D-DCT method

Can¹,

Karaca²,

Urhan³

et al. 2020

Pamukkale J Eng Sci

View full text Add to dashboard Cite

Hiperspektral görüntüleme, farklı uygulama alanlarındaki kullanımı ile son yıllarda oldukça popüler bir konu haline gelmiştir. Yüksek depolama alanlarına ihtiyaç duyan hiperspektral görüntülerin yüksek verim ve kalite ile sıkıştırılması gerekmektedir. Bu çalışmada, hiperspektral görüntülerin kayıplı sıkıştırılması için otomatik uyarlamalı ışıklılık dönüşümü ve üç-boyutlu ayrık kosinüs dönüşümünü (3D-DCT) kullanan özgün bir yöntem önerilmektedir. Önerilen yöntemde ilk olarak hiperspektral verideki spektral bantlar gruplanmış ve ön işlem olarak otomatik uyarlamalı ışıklılık dönüşümü uygulanmıştır. Elde edilen her bant grubu ayrık kosinüs dönüşümü ve sonrasında Huffman kodlama kullanılarak sıkıştırılmıştır. Önerilen ışıklılık dönüşümünün amacı, bir grup içindeki bant imgeleri arasındaki ışıklılık ve karşıtlık farklılıklarını azaltarak sıkıştırma performansının arttırılmasını sağlamaktır. Deneysel sonuçlarda, Cuprite, Moffett Field, Jasper Ridge ve Pavia University hiperspektral görüntüleri üzerinde önerilen yöntem, ışıklılık dönüşümünün farklı versiyonları ile karşılaştırılmıştır. Karşılaştırma sinyal-gürültü oranı ve ortalama spektral açı uzaklığı gibi ölçütler kullanılarak yapılmıştır. Bunun yanında, sıkıştırılan verideki anomali ve hedef tespiti başarımları da karşılaştırılmıştır. Önerilen yöntemin, 3D-DCT sıkıştırma performansını özellikle düşük bit oranlarında ortalama %40 oranına kadar arttırdığı gösterilmiştir. In recent years, hyperspectral imaging has become a very popular subject with its use in different application areas. Hyperspectral images that require high storage areas need to be compressed with high efficiency and quality. In this study, a novel method that uses automatic adaptive luminance transform and three-dimensional discrete cosine transform (3D-DCT) for lossy compression of hyperspectral images is proposed. Firstly, spectral bands in hyperspectral image are grouped and automatic adaptive luminance transform is performed as a preprocessing stage in the proposed method. Each group is compressed by using DCT and Huffman encoding. The aim of the proposed luminance transform is to increase compression performance by decreasing luminance and contrast differences between band images in a group. In the experimental results, the proposed method and different versions of luminance transform are compared on Cuprite, Moffet Field, Jasper Ridge and Pavia University hyperspectral images. Comparison is carried out using signal-to-noise ratio and average spectral distance metrics. Besides, anomaly and target detection performances are also compared for compressed images. The proposed method has been shown to increase compression performance of 3D-DCT up to an average of 40% rate, especially at low bit rates.

show abstract

“…Nowadays, with the development of technologies for airborne systems and space-borne remote-sensing sensors, the ability to obtain data in both high spatial and high spectral resolution is increasingly feasible (Scheffler & Karrasch, 2013). This leads to the use of remote-sensing satellites in various areas of Earth science studies, including object detection Cheng, Zhou, & Han, 2016), image classification (Cheng, Han, Zhou, & Guo, 2014;Cheng et al, 2015;Hassanzadeh & Karami, 2016;Raczko & Zagajewski, 2017), anomaly detection (AD) (L. Zhang & Zhao, 2017) and CD (Shah- Hosseini, Homayouni, & Safari, 2015;K. Wu et al, 2015).…”

Section: Introductionmentioning

confidence: 99%

A new land-cover match-based change detection for hyperspectral imagery

Seydi

Hasanlou

2017

European Journal of Remote Sensing

View full text Add to dashboard Cite

The presence of phenomena such as earthquakes, floods and artificial human activities causes changes on the Earth's surface. Change detection (CD) is an essential tool for the monitoring and managing of resources on local and global scales. Hyperspectral imagery can provide more detailed results for detecting changes in land-cover types. The main objective of this paper is to present a new, supervised CD method by combining similarity-based and distance-based methods to increase the efficiency of already existing CD approaches. The proposed method applies in two phases and uses three different algorithms, including image differencing, modified Z-score analysis and spectral angle mapper. The efficiency of the presented method is evaluated using Hyperion multi-temporal hyperspectral imagery. The receiver-operating characteristic curve index is used for assessing and comparing the results. The results clearly demonstrate the superiority of the proposed method for the detection and production of more accurate change maps. Furthermore, the proposed method is also able to detect changes with an accuracy of more than 96%, a false alarm rate lower than 0.03 and an area under the curve of about 0.986 in overall comparison to other conventional CD techniques. In addition, this method achieved an optimal threshold value with more rapid convergence.

show abstract

Compression and noise reduction of hyperspectral images using non-negative tensor decomposition and compressed sensing

Cited by 12 publications

References 16 publications

An Efficient Compressive Hyperspectral Imaging Algorithm Based on Sequential Computations of Alternating Least Squares

An Efficient Compressive Hyperspectral Imaging Algorithm Based on Sequential Computations of Alternating Least Squares

Compression of hyperspectral images using automatic adaptive luminance transform and 3D-DCT method

A new land-cover match-based change detection for hyperspectral imagery

Contact Info

Product

Resources

About