T. Kusaka

Radiometric correction for atmospheric and topographic effects on Landsat MSS images

Abstract Raw radiance data measured by a Landsat satellite includes undesirable atmospheric and topographic effects. The correction to allow for these effects is important in order to increase the accuracy of classification. In this paper we show how these effects on the Landsat MSS data are estimated, allowing for the transfer theory of radiation in the atmosphere-ground system. First, we present the theoretical basis of the atmospheric effect correction system (AECS) which corrects for atmospheric effects on the Landsat MSS data over a flat terrain. Then we propose a simple radiometric correction method which can remove both the atmospheric and topographic effects from the Landsat MSS data over a rugged terrain. We have applied it to a mountainous test area with known digital terrain data and obtained satisfactory results. In our analyses, the values of relevant atmospheric parameters such as the optical thickness, the single scattering albedo and the turbidily factor of the atmosphere were adopted from...

Region Extraction in SPOT Data

Abstract This paper presents a method for segmenting SPOT HRV data using color edge points to separate small regions with nearly constant color. Region segmentation using edge points has not been successful because edge detection operators yield the partially enclosed regions. The present method uses isograms of equal distances from the boundary points between the partially enclosed regions. The small enclosed regions (primitive regions) are separated, based on the local isogram maxima and their values. After SPOT data have been segmented, spectral and spatial features such as color, vegetation index, size and shape are extracted for each primitive region. It is shown that the primitive regions corresponding to typical land cover types (rice field, urban area, residential area, forest, highway and river) are successfully characterized by spectral and spatial features.

Atmospheric and topographic correction of satellite data over mountainous terrain

An analytic model for the radiance applicable to the remote sensing of mountainous terrain is described. The atmosphere is treated as a multiple scattering, horizontally uniform layer bounded by non-homogeneous Lambertian surface. This radiance model includes the radiation arising from adjacent terrain outside the instantaneous field of view (IFOV) of the sensor. The authors take occlusion by adjacent terrain into account in estimating the diffuse sky irradiance on the target. Furthermore, the correction algorithm for removing atmospheric and topographic effects from high-resolution satellite images is formulated, and the image correction for the Landsat TM image is performed.<<ETX>>

Extraction of Landslide Areas Using Spatial Features of Topographic Basins

Spatial features of watersheds are examined to clarify characteristic properties in areas where landslides may occur. In this study, first, a method for delineating watersheds from the digital terrain data is described. Next, the area (A), mean slope (S) and form factor (F=A/L2) of watersheds are estimated. Here, L is the length of the main stream in a watershed. As a result, it is shown that topographic features in landslide areas are successfully characterized by a linear combination of the normalized values of A, S and F.

Extraction of characteristic properties in landslide areas using thematic map data and surface temperature

It is well known that in the Noto area in Hokuriku district of Japan, landslides occur in all seasons. Most small-scale landslides, which the authors have investigated, occur near streams in valleys with a gentle slope. Moreover, it is said that underground-water supplied by rainfall and snowfall strongly relates to these landslides. The objective of this study is to find areas where landslides may occur in the near future by using infrared band data and visible band data obtained from a Landsat TM sensor with Thematic-Mapper data related to landslide areas. As a result, at many landslide segments in investigated areas, the authors found that the surface temperature in landslide areas is higher than in non-landslide areas.<<ETX>>

Extraction of topographic features in landslide areas from ERS-1 SAR data

An analysis has been undertaken of characteristic features of the ERS-1 SAR image in areas where landslides may occur. First, statistical quantities of SAR data corresponding to landslide areas and non-landslide areas are computed for the study site (the Noto peninsula in Japan). As a result, it is shown that the difference of mean values of gray levels between landslide areas and non-landslide areas is of no significance. Moreover, the drainage system is extracted on the basis of the assumption that some streams will be defined by the dark strips caused by the combined effects of shadow and open water. It is found that the extracted drainage courses correspond to the dark strips that are visible on the SAR image as valleys.<<ETX>>

Signature Variations Due To Atmospheric And Topographic Effects On Satellite MSS Data Over Rugged Terrain

A model for the radiance applicable to remote sensing of mountainous terrain is described. T his radiance model includes the r adiation arising from adjacent terrain outside t he instantaneous field of view (IFOV) of the sensor. The atmosphere is treated as the optically t hin, horizontally uniform slab layer bounded by non-homogeneous Lambertian surface. The results from the r adiance equation show that the sky irradiance component on the target due to multiple reflections by adjacent terrain is of no significance. It is also shown that the path radiance reflected by the background is important in the poorly illuminated region.

Extraction of landslide areas using satellite remote sensing and GIS technology

The objective of this study is to find areas where landslides may occur in the near future by using satellite remote sensing and thematic-map data, also using GIS (Geographic Information System) techniques. The authors considered how the combination of GIS with remote sensing data and thematic map data may facilitate greatly the assessment and estimation of regional landslide hazards. They extracted inclination angles, inclination aspects, and ridge and valley lines from DTM data originally made for landslide areas. The watersheds in the study areas were extracted by GIS techniques. Criteria to predict landslide areas are proposed.

Atmospheric correction algorithm for remote sensing data with multi-viewing angles

The atmospheric correction algorithms, developed previously for the conventional scanned image data, like Landsat MSS and TM data, are no longer valid because the viewing angle direction at one target pixel is different from that at other pixel in the frame image. The authors present the theoretical basis of a new atmospheric correction algorithm for the frame image data.<<ETX>>

Inference of landslide areas using spatial features and surface temperature of watersheds

An inference of landslide areas using digital elevation data and Landsat TM band 6 data has been performed on the basis of the assumption that the occurrence of landslides is closely related to the amount of underground-water and the topographic features of watersheds. It is shown that it is possible to distinguish between dangerous landslide areas and non-landslide areas by using spatial features of watersheds, such as the area, mean slop and shape factor, and the ground surface temperature obtained from Landsat TM data.