Fuzzy classification of Earth terrain covers using multi-look polarimetric SAR image data

Du, Li-Jen; Lee, J.S.; Mango, S.

doi:10.1109/igarss.1993.322304

Cited by 13 publications

(11 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the polarimetric covariance matrix, distance measures between an unknown pixel and a class center are defined [14], [15]. The POLSAR image is thus classified using supervised or unsupervised classifiers such as ISODATA, Neural network [16] and fuzzy -mean [17]; and 3) advanced methods that combine physical scattering decomposition and statistical classification. The entropy/alpha-Wishart classifier [18] is one of the most widely applied approaches in this category, which initializes training sets by entropy/alpha decomposition [19] and performs iterations of maximum likelihood classification base on the complex Wishart distribution of refined training sets [3], [20], [21].…”

Section: Improved Building Extraction With Integratedmentioning

confidence: 99%

Improved Building Extraction With Integrated Decomposition of Time-Frequency and Entropy-Alpha Using Polarimetric SAR Data

Deng

Wang

2014

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Building extraction is one of the primary applications of urban remote sensing. Polarimetric synthetic aperture radar (POLSAR), with its all-weather day and night imaging, canopy penetration and full polarimetric information, provides a unique way to detecting and characterizing urban areas. In this study, the time-frequency decomposition technique and the entropy/alpha-Wishart classifier were integrated to improve building extraction. The entropy/alpha-Wishart classifier was able to extract ortho-oriented buildings. After time-frequency transformation, the variation of entropy, alpha, anisotropy differs for objects with different scattering mechanisms, and the alpha angle of subaperture images was optimal in delineating slant-oriented buildings. A comparison between the integrated approach and the conventional entropy/alpha-Wishart classifier was performed on both C-and L-band NASA/JPL AIRSAR datasets. The overall accuracy and kappa value of the integrated approach was about 20% higher than that of the entropy/alpha-Wishart classifier. The C-band output tends to show more detailed scattering properties whereas the extracted buildings from the L-band image reveal better overall visual results.

show abstract

Section: Improved Building Extraction With Integratedmentioning

confidence: 99%

Improved Building Extraction With Integrated Decomposition of Time-Frequency and Entropy-Alpha Using Polarimetric SAR Data

Deng

Wang

2014

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

show abstract

“…In the case of the simplified Wishart-distribution of equation (8), equation (11) and equation (13) together lead to an improvement of the log-likelihood with respect to AE ðkÞ j and assignments y (Bezdek 1981, Du and Lee 1996, Chen et al 2003. To reach a local maximum, the expectation and maximization steps are carried out iteratively until a certain termination criterion is met.…”

Section: Classification Via Expectation Maximizationmentioning

confidence: 96%

“…Alternatives are the convergence of the class centres themselves (Du and Lee 1996), the percentage of pixels changing their most likely class between iterations falling below a certain threshold, or simply a fixed number of iterations. Figure 1 shows a quantitative comparison of the expectation classification technique with threshold classification of the H-a feature space (Cloude and Pottier 1997) and Wishart k-means classification (Lee et al 1999).…”

Section: Classification Via Expectation Maximizationmentioning

confidence: 99%

Classifying polarimetric SAR data by combining expectation methods with spatial context

Reigber

Jäger

Neumann

et al. 2010

International Journal of Remote Sensing

View full text Add to dashboard Cite

show abstract

“…However, the mixed-pixel problem was not considered, even though fully polarimetric features were used. Du and Lee [11], [22] integrated a complex Wishart distribution and fuzzy c-means (FCM) clustering in an unsupervised technique. Neural networks has the capability of adopting varied types of feature vector while retaining highly flexible in terms of their structure and learning algorithms [2], [3], [5]- [10], [12].…”

Section: Introductionmentioning

confidence: 99%

The use of fully polarimetric information for the fuzzy neural classification of SAR images

Chen

Lee

2003

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

This paper presents a method, based on a fuzzy neural network, that uses fully polarimetric information for terrain and land-use classification of synthetic aperture radar (SAR) image. The proposed approach makes use of statistical properties of polarimetric data, and takes advantage of a fuzzy neural network. A distance measure, based on a complex Wishart distribution, is applied using the fuzzy c-means clustering algorithm, and the clustering result is then incorporated into the neural network. Instead of preselecting the polarization channels to form a feature vector, all elements of the polarimetric covariance matrix serve as the target feature vector as inputs to the neural network. It is thus expected that the neural network will include fully polarimetric backscattering information for image classification. With the generalization, adaptation, and other capabilities of the neural network, information contained in the covariance matrix, such as the amplitude, the phase difference, the degree of polarization, etc., can be fully explored. A test image, acquired by the Jet Propulsion Laboratory Airborne SAR (AIRSAR) system, is used to demonstrate the advantages of the proposed method. It is shown that the proposed approach can greatly enhance the adaptability and the flexibility giving fully polarimetric SAR for terrain cover classification. The integration of fuzzy c-means (FCM) and fast generalization dynamic learning neural network (DLNN) capabilities makes the proposed algorithm an attractive and alternative method for polarimetric SAR classification.

show abstract

Fuzzy classification of Earth terrain covers using multi-look polarimetric SAR image data

Cited by 13 publications

References 6 publications

Improved Building Extraction With Integrated Decomposition of Time-Frequency and Entropy-Alpha Using Polarimetric SAR Data

Improved Building Extraction With Integrated Decomposition of Time-Frequency and Entropy-Alpha Using Polarimetric SAR Data

Classifying polarimetric SAR data by combining expectation methods with spatial context

The use of fully polarimetric information for the fuzzy neural classification of SAR images

Contact Info

Product

Resources

About