Based on the measurements with HHRR (Hand Held Rationing Radiometer) which has a similar spectral responsivity to that of Landsat Thematic Mapper, characteristics of spectral reflectivity of several non-metallic ores and altered rocks collected in Yangsan-Dong Rae Faults in South Korea are made. The collection includes pyrophyllite ores, andesites, silicified sedimentary rocks, granite and granodiorite. The analyses of the measured data indicate following features.
1) Among the samples pyrophyllite ore has the highest reflectance in the spectral range corresponding to Landsat TM Band-1 through 5. The reflectance of pyrophyllite ore depends on quality with higher reflectance for higher quality.
2) The altered rock has higher reflectance than the unaltered rock.
3) Sharp drop of reflectance at 2.22 pm region corresponding to Band-7 of Landsat TM is observed for pyrophyllite which is considered due to the effect of OH. All the samples show low reflectance at Band-4 which is considered due to the effect of oxidization.
4) The band ratio of Band-7/Band-5 and Band-7/Band-4 are effective to discriminate pyrophyllite from other rocks.

View full abstract

The effectiveness of the radar remote sensing is investigated. As is well known, distortions of radar images are influenced by radar and terrain parameters. Thus for the geological applications, relations between ray-orientation and surface geometry are important key controlling items.
This paper describes the applicability of computer-simulated, pseudo, images to assess such parameters. To simulate images, the authors assumed a very simple wave scattering model. In spite of the model simplicity, quite a remarkable coincidence is obtained between actual and pseudo images, thus proving the adequacy of the theory.
Terrain distortions, layover, foreshortening, backelongation and shadowing as well are quantitatively evaluated changing the angles of the incidence. The horizontal look-direction effects also investigated using 2-D Fourier spectra of the images.
Thus obtained results are, 1) the radar scattering is simulated by very simple models, 2) lineaments perpendicular to the horizontal direction of the radar beam applied are selectively enhanced, 3) for the terrain mapping, relatively large incident
angles, i. e. at least 45° or larger, preferably 60° are required.

View full abstract

Various types of "second generation" SAR with multi-sensing mode, replacing "classic" SAR like one on Seasat have been proposed to meet many user requirements.
In this paper, a design method of SAR with variable sensor system parameters utilizing unfocused SAR principal positively is proposed to make variable off-nadir angle observation with constant high spatial resolution in costant swath width, available with one SAR system. And inherent technological problem on the ambiguity of SAR with variable off-nadir angle are discussed.

View full abstract

This paper describes the geographical investigation of land cover classification and its changes in the central part of the Kanto Plain using Landsat data. At first, land cover classification was observed by visual interpretation of color composite images, the principal component images from multi-temporal Landsat images, and land cover classification analysis using Landsat data. Secondly, geographical distribution and detailed change of land cover were investigated by comparing land cover trend surfaces from Landsat data with land productivity trend surfaces from agricultural grid data. Furthermore, the validity of trend surface analysis and comparison of Landsat data with digital national land information were investigated.
Conclusions of this study are shown as follows.
1) The correspondence of land cover conditions with geomorphological conditions was shown by the visual interpretation of Landsat false color composite images and a TM principal component image.
2) In the land cover classification using Landsat data, the increase of urbanization land cover was shown in the southern part of study area and along the railroad.
3) Land cover change images and urbanization trend surfaces indicated that the residential development of the 1970's increased on terrace and the large scale residential development of the 1980's increased on lowland.
4) By using the trend surface analysis, the condition of urbanization was made clear and the validity of trend surface fitting for land cover classification was confirmed.
5) The land cover information from Landsat data is comparable with the digital national land information and is very useful in regional analysis of quantitative geography.

View full abstract

Concrete crack inspection has become very important in recent years, because the deterioration of cocrete structure is progressing. This kind of inspection has been the visual inspection for the most part so far. The authors have develped a new method for such inspection using an image processing system assembled from a personal computer and image processing softwares.
The procedure is firstly to take the surface condition of cocrete structure by video camera into video tape, and then to analyze the image for concrete crack inspection. The authors had two applications. One is an analysis of the progress of cocrete breaking of test piece in the laboratory. Another is an analysis of condition of the cracks upon the wall surface of an used building. Two results are obtained;
(1) Histogram pattern of crack existing area is flatter than that of non-crack existing area. In addition, a new mound is born at crack generation and grows with the increment of crack area.
(2) The above result was same in the test piece of the laboratory and in the wall surface of the build-ing.

View full abstract

Recent years, some countries launched there own satellite successfully, and it becomes possible to
collect the earth surface information by different remote sensors. Merging the image data collected by different sensors is becoming an important subject of digital processing. In this study, two data sets with different characteristics were digitally merged, and a single data set, which contains information from both sets, was generated. Landsat TM (Thematic Mapper) data, which have seven spectral bands, was. selected for the spectral information, and Spot HRV (High Resolution Visible Infrared Radiometer) pancromatic data, which spatial resolution is 10 m, was selected for the spatial information.
The Landsat TM data were digitally enlarged by resampling method to produce the same resolution of
Spot HRV data. Registration accuracy of both data set depending on affine transformation using 24 ground control points, was approximately 6.3 meters in the column and 3.4 meters in the line. This
results are because of the selection of test site as a small and low undulation area.
A linear combination method is used for digital merging of both data, and the output image attains
to a successful level in the spatial and spectral resolution.
by S. Tanaka and T. Sugimura

View full abstract