Bruce Forster

An examination of some problems and solutions in monitoring urban areas from satellite platforms

Abstract A discussion of the benefits of satellite remote sensing to urban studies is followed by a consideration of resolution requirements and associated problems. Problems include loss of contextual clues for interpretation, heterogeneity of cover surfaces, temporal differences in atmospheric effects and registration of different scenes. A number of solutions to these problems are considered, such as the use of reference surfaces for atmospheric corrections and the use of targetted control points for temporal change monitoring. The specific advantages of high-resolution satellite data are particularly considered

Radar backscatter analysis for urban environments

Abstract A full understanding of radar backscatter from urban areas is necessary in order to develop a robust methodology for monitoring and classifying urban characteristics using remotely sensed Synthetic Aperture Radar images.This paper examines the dominant backscattering mechanisms such as single bounce from roofs, double bounce from wall-ground structures and possibly triple bounce from wall-wall-ground structures, and their relative contributions to the backscatter. With the use of quad-polarized image data such as those acquired by the NASA/JPL AirSAR system, the backscatter can be decomposed into components caused by different backscattering mechanisms, offering a promise for urban monitoring and classification.

Derivation of atmospheric correction procedures for LANDSAT MSS with particular reference to urban data

Abstract The complexities and detail of urban scenes make it imperative that atmospheric effects are removed from satellite remotely sensed data prior to analysis. A study of atmospheric theory allows a simplified procedure to be developed for correction of multitemporal LANDSAT MSS scenes. Examples in the visible and near infrared, from a summer and winter scene, illustrate how disparate count values can be brought to good agreement as percentage reflectance.

Segmentation and classification of vegetated areas using polarimetric SAR image data

Classification of radar images based on the information provided by individual pixels cannot generally produce satisfactory results due to speckle. The classification based on area analysis is therefore expected to be more accurate, as a uniform area, which usually consists of multipixels, provides reliable measurement statistics and texture characteristics. However, the area analysis requires partitions of uniform areas to be performed first. In this paper, an approach to the classification of radar images is developed based on two steps. First an image is partitioned into uniform areas (segments), and then these segments are classified. Both segmentation and classification are achieved by using the Gaussian Markov random field model. Test images are classified to demonstrate the method.

A new decomposition of radar polarization signatures

A method for the decomposition of radar polarization signatures is developed. The polarization backscattering model is assumed to consist of odd, double, Bragg, and cross backscattering components, and the Mueller matrix is the sum of the Mueller matrices of these four scattering mechanisms. The technique of least squares (LS) is then used to find the best combination of these four components. Using NASA/Jet Propulsion Laboratory (JPL) AirSAR data, the results of the decomposition agree with the general understanding of radar backscatter. In most cases, the accuracy of the decomposition is more than 95% for linear polarizations and more than 85% for any other polarizations.

Comparison of radar image segmentation by Gaussian- and Gamma-Markov random field models

This paper compares segmentation results of synthetic aperture radar (SAR) images using Gaussian-Markov random field (MRF) and Gamma-MRF models. A Gamma distribution function is more accurate and proper to trace the multilook SAR intensity data distribution. However, it is found that, at least from examples used in the paper, when the distribution function is incorporated with the MRF model to implement SAR image segmentation, the Gamma-MRF model is not necessarily shown to be superior to the Gaussian-MRF model. Occasionally the Gamma-MRF model wrongly merges a few small segments, suggesting that the Gaussian-MRF model might be more stable and reliable.

Remote sensing of sea water quality parameters using Landsat-TM

Abstract This paper describes an experiment where sea water quality parameters were determined using data from the Landsat Thematic Mapper (TM) satellite remote sensing system over a coastal sewage outfall area. The parameters determined included turbidity, chlorophyll-a, chlorophyll-i, phaeopigment and total pigment. The area investigated was a sewage outfall site off the North Head of Sydney Harbour, Australia. The method used multiple regression to relate site sampled parameters to digital Landsat-TM data, and the results verified using data not used in the regression. Multiple correlation coefficients in excess of R = 0. 9 resulted from the regression analysis, which used Landsat-TM variables in a Chebyshev Series form. However due to the limited number of ocean samples only the results for turbidity were considered significant. Nevertheless a satisfactory methodology is proposed.

A review of SAR speckle filters: texture restoration and preservation

Speckle filtering is probably one of the important steps in post SAR (synthetic aperture radar) image processing. Speckle appears in all SAR products including single-polarimetric, multipolarimetric and interferometric image data. Many speckle filtering algorithms have been proposed aimed at speckle reduction. Texture, revealing the spatial-dependent information in images, may be as important as other polarimetric measures. Such information is not expected being distorted and degraded during speckle suppression. This paper examines various speckle filters including the Lee, Frost, Kuan and other filters in terms of texture preservation. The first and second order texture measures are used in the assessment.

Coefficient of variation as a measure of urban spatial attributes, using SPOT HRV and landsat TM data

Abstract The variation of radiometric response over urban residential areas, as recorded by a sensor, is considered to be a function of the size and spatial distribution of housing units and the pixel size of the remote sensing system. Using data from thirty residential sites over the Sydney metropolitan area, and SPOT HRV and Landsat TM data, the relation between radiometric variability and the urban spatial measures of house size and housing density were analysed. Preliminary results show that the coefficient of variation is both a function of housing density and house size, for combined multiple-sensor data sets of different pixel sizes, and strongly related to house size for different classes of housing density using a single band of SPOT HRV XS-mode data.

Hail storm vulnerability assessment by using hyperspectral remote sensing and GIS techniques

Describes a methodology which uses hyperspectral remote sensing data and cartographic GIS data for assessing and mapping the vulnerability from hail storm disaster in Sydney, Australia. By collecting urban surface material samples, mainly comprising roof types, a spectral library was created under artificial illumination using a spectroradiometer. The reference spectra was resampled and forced to the Hymap image spectra by the empirical line calibration method. A spatial distribution map showing vulnerable areas to hail storm hazard was generated by the spectral angle mapper method in near real time. This database upon integration with available cartographic data has the potential to develop into a powerful decision support system for emergency organisations. Results out of this study show great potential in hyperspectral applications and mapping for emergency planning.