A detailed comparison of backpropagation neural network and maximum-likelihood classifiers for urban land use classification

Paola, J D; Schowengerdt, R A

doi:10.1109/36.406684

Title: A detailed comparison of backpropagation neural network and maximum-likelihood classifiers for urban land use classification

Full Record
Other Related Research

Abstract

A detailed comparison of the backpropagation neural network and maximum-likelihood classifiers for urban land use classification is presented in this paper. Landsat Thematic Mapper images of Tucson, Arizona, and Oakland, California, were used for this comparison. For the Tucson image, the percentage of matching pixels in the two classification maps was only 64.5%, while for the Oakland image it was 83.3%. Although the test site accuracies of the two Tucson maps were similar,the map produced by the neural network was visually more accurate; this difference is explained by examining class regions and density plots in the decision space and the continuous likelihood values produced by both classifiers. For the Oakland scene, the two maps were visually and numerically similar, although the neural network was superior in suppression of mixed pixel classification errors. From this analysis, the authors conclude that the neural network is more robust to training site heterogeneity and the use of class labels for land use that are mixtures of land cover spectral signatures. The differences between the two algorithms may be viewed, in part, as the differences between nonparametric (neural network) and parametric (maximum-likelihood) classifiers. Computationally, the backpropagation neural network is at a serious disadvantage to maximum-likelihood,more »« less

Authors:

Paola, J D; Schowengerdt, R A ^[1]

Univ. of Arizona, Tucson, AZ (United States). Dept. of Electrical and Computer Engineering

Publication Date:: Sat Jul 01 00:00:00 EDT 1995

OSTI Identifier:: 136699

Resource Type:: Journal Article

Journal Name:: IEEE Transactions on Geoscience and Remote Sensing

Additional Journal Information:: Journal Volume: 33; Journal Issue: 4; Other Information: PBD: Jul 1995

Country of Publication:: United States

Language:: English

Subject:: 54 ENVIRONMENTAL SCIENCES; URBAN AREAS; LAND USE; AERIAL SURVEYING; IMAGE PROCESSING; NEURAL NETWORKS; CLASSIFICATION; GROUND COVER; REMOTE SENSING

Citation Formats


                    Paola, J D, and Schowengerdt, R A. A detailed comparison of backpropagation neural network and maximum-likelihood classifiers for urban land use classification.  United States: N. p., 1995. 
        Web.  doi:10.1109/36.406684.

Copy to clipboard


                    Paola, J D, & Schowengerdt, R A. A detailed comparison of backpropagation neural network and maximum-likelihood classifiers for urban land use classification.  United States.  https://doi.org/10.1109/36.406684

Copy to clipboard


                    Paola, J D, and Schowengerdt, R A. 1995.  
        "A detailed comparison of backpropagation neural network and maximum-likelihood classifiers for urban land use classification".  United States.  https://doi.org/10.1109/36.406684.

Copy to clipboard


                    
@article{osti_136699,

  title        = {A detailed comparison of backpropagation neural network and maximum-likelihood classifiers for urban land use classification},

  author       = {Paola, J D and Schowengerdt, R A},

  abstractNote = {A detailed comparison of the backpropagation neural network and maximum-likelihood classifiers for urban land use classification is presented in this paper. Landsat Thematic Mapper images of Tucson, Arizona, and Oakland, California, were used for this comparison. For the Tucson image, the percentage of matching pixels in the two classification maps was only 64.5%, while for the Oakland image it was 83.3%. Although the test site accuracies of the two Tucson maps were similar,the map produced by the neural network was visually more accurate; this difference is explained by examining class regions and density plots in the decision space and the continuous likelihood values produced by both classifiers. For the Oakland scene, the two maps were visually and numerically similar, although the neural network was superior in suppression of mixed pixel classification errors. From this analysis, the authors conclude that the neural network is more robust to training site heterogeneity and the use of class labels for land use that are mixtures of land cover spectral signatures. The differences between the two algorithms may be viewed, in part, as the differences between nonparametric (neural network) and parametric (maximum-likelihood) classifiers. Computationally, the backpropagation neural network is at a serious disadvantage to maximum-likelihood, taking nearly an order of magnitude more computing time when implemented n a serial workstation.},

  doi          = {10.1109/36.406684},

  url          = {https://www.osti.gov/biblio/136699},
  journal      = {IEEE Transactions on Geoscience and Remote Sensing},
number       = 4,

  volume       = 33,

  place        = {United States},

  year         = {Sat Jul 01 00:00:00 EDT 1995},

  month        = {Sat Jul 01 00:00:00 EDT 1995}

}

Copy to clipboard

Journal Article:

https://doi.org/10.1109/36.406684

Other availability

Find in Google Scholar

Search WorldCat to find libraries that may hold this journal

Save / Share:

Export Metadata

Save to My Library

Similar records in OSTI.GOV collections:

Selection of optimal textural features for maximum likelihood image classification

Technical Report Rosenblum, W.; Salvaggio, C.; Schott, J.

The classification or segmentation of images into land cover or object class is one of the fundamental remote sensing/image processing tasks. The classification techniques have reached a high level of maturity in the forms of statistical pattern recognition techniques applied to multispectral images. According to this approach each pixel has a spectral vector associated with it and pixels are segmented into the class they most closely resemble spectrally. While these techniques are reasonably successful they do not take advantage of the brightness patterns within a class or at object boundaries nor are they readily applicable to monochrome images or highlymore »« less
https://doi.org/10.2172/5098367

Full Text Available
Data classification, visualization, and enhancement using n-dimensional probability density functions (nPDF). AVIRIS, TIMS, TM, and geophysical applications

Journal Article Cetin, H; Warner, T; Levandowski, D - Photogrammetric Engineering and Remote Sensing

The n-Dimensional Probability Density Functions (nPDF) approach is a user-interactive image analysis technique which overcomes many of the inherent limitations of traditional classifiers. In this paper we illustrate the applications of nPDF analysis in three broad areas: data visualization, enhancement, and classification. For data visualization, nPDF provides a method for transforming multiple bands of data in a predictable and scene-independent way. These transformations may be designed so as to enhance a particular cover type, or to give the best visual representation of the multi-band image data. These approaches are illustrated with the enhancement of hydrothermally altered areas in Thematic Mappermore »« less
star Miner: A suit of classifiers for spatial, temporal, ancillary, and remote sensing data mining

Conference Vatsavai, Raju; Shekhar, Shashi; Burk, Thomas; ...

Thematic classification of multi-spectral remotely sensed imagery for large geographic regions requires complex algorithms and feature selection techniques. Traditional statistical classifiers rely exclusively on spectral characteristics, but thematic classes are often spectrally overlapping. The spectral response distributions of thematic classes are dependent on many factors including terrain, slope, aspect, soil type, and atmospheric conditions present during the image acquisition. With the availability of geo-spatial databases, it is possible to exploit the knowledge derived from these ancillary geo-spatial databases to improve the classification accuracies. However, it is not easy to incorporate this additional knowledge into traditional statistical classification methods. On themore »« less
star Miner: A suit of classifiers for spatial, temporal, ancillary, and remote sensing data mining

Journal Article Burk, Thomas; Bhaduri, Budhendra - Proceedings of the Fifth International Conference on Information

Thematic classification of multi-spectral remotely sensed imagery for large geographic regions requires complex algorithms and feature selection techniques. Traditional statistical classifiers rely exclusively on spectral characteristics, but thematic classes are often spectrally overlapping. The spectral response distributions of thematic classes are dependent on many factors including terrain, slope, aspect, soil type, and atmospheric conditions present during the image acquisition. With the availability of geo-spatial databases, it is possible to exploit the knowledge derived from these ancillary geo-spatial databases to improve the classification accuracies. However, it is not easy to incorporate this additional knowledge into traditional statistical classification methods. On themore »« less
Classifying Radio Galaxies with the Convolutional Neural Network

Journal Article Aniyan, A.; Thorat, K. - Astrophysical Journal, Supplement Series

We present the application of a deep machine learning technique to classify radio images of extended sources on a morphological basis using convolutional neural networks (CNN). In this study, we have taken the case of the Fanaroff–Riley (FR) class of radio galaxies as well as radio galaxies with bent-tailed morphology. We have used archival data from the Very Large Array (VLA)—Faint Images of the Radio Sky at Twenty Centimeters survey and existing visually classified samples available in the literature to train a neural network for morphological classification of these categories of radio sources. Our training sample size for each of these categoriesmore »« less
https://doi.org/10.3847/1538-4365/AA7333

Similar Records