Earth's surface lights observed from satellite indicate various aspects about artificial and natural phenomena. Major patterns of light distribution are caused from human activities in cities, forest, farmland, ocean, oil field, etc. Global assessment of light distribution is important for us to investigate the state of natural environment on the Earth's surface and its change with human activity. First, gathering 38 DMSP night light photographs in 1984 to 1988, a DMSP global nighttime image was created. Source data were DMSP Block-5D satellite films stored at Colorado University, a branch of World Data Center A. Thanks to the staffs of the University, cloud free or almost less data could be obtained. Second, a compilation process was taken by A/D conversion from photographs to digital data. After taking two processes, radiometric and geometric correction, a global scale image was produced. Ground control points were used for regional film orientation, and resampling and mosaicking processes were taken. Our final approach was an interpretation of geographical meaning of the night light in global scale, specially with human activities. We summarized our idea as follows. 1) DMSP nighttime images with selected GCPs about 14 to 41 points after radiometric correction were oriented to a plate carree coordinates by three cubic polynomials. The registration accuracy was about 1.0 pixels and 1.0 lines. 2) A DMSP global nighttime image could be produced by resampling and mosaicking operations on plate carree coordinates. It was proved that the geographical position and the range of light distribution could be investigated in detail on the image combining both of the GVI data and the geographical information. 3) Relationship between light distribution and the human activity was investigated. 4) It was clearly shown that the distribution patterns of light based on sifting cultivation were located in forest of mountainous district and savanna zone. 5) It was searched that the both fishing fire and the flare of oil field had the regional feature. 6) The light distribution pattern on the nighttime image is a description of natural energy consumption. It is thought that the influence of human activity for the natural environment can be investigated from this new viewpoint, if detailed analysis of the nighttime image could continuously be done with the other auxiliary sources.
The nPDF (n-dimensional Probability Density Function) is an algorithm for displaying, analyzing and classifying data (H. Cetin, 1990). The nPDF algorithm is useful for multi-dimensional data transformation and reduction. Furthermore the nPDF plots provide a clear perspective of the data distributions. In this study, we newly define nPDF features as the spatial information. This paper discusses whether the classification accuracy is improved or not with Maximum Likelihood Classification (MLC) in combination with the spectral information and the nPDF features. Three cases, for TM and HRV data respectively, were executed as follows : 1) Only using the spectral information, 2) Using nPDF features in addition to the spectral information, and 3) Using the enhancement image with Laplacian operator considering the direction of 45 degrees which was reported as one of the useful spatial information for MLC (T. Tiyip, 1991) The results of this study are as follows : 1) In case of using nPDF features, the improvement of the classification accuracy is confirmed for both TM and HRV data. 2) Furthermore, it was found that nPDF features as a spatial information is more useful for improving the classification accuracy than the enhancement image. 3) As a result of measurement of CPU time (computing time by using IBM 9121 : Type 320), it takes only about 36 seconds to make the nPDF feature (500×500 pixels). The transaction of data reduction and transformation using nPDF algorithm is fast and dose not require much memory for calculation ; accordingly the practicality should be assured for multi-spectral classification using nPDF features as a spatial information.
It is well known that the radar backscattering characteristics of the tropical rain forest show very small dependencies on the incidence angle for wide range of the radar frequencies through several experiments. Based on this fact, an estimation of the SAR antenna elevation pattern for the JERS-1 satellite was conducted through two steps, (1) the statistical screening of the SAR data over the tropical rain forest to obtain the same characterized uniform images, and (2) the application of the correlated SAR image model including the noise bias, which should be considered for more correct estimation of the antenna beam width, to the screened SAR images by means of the least square method. Finally, one way antenna elevation pattern was obtained to 5.44 degree in terms of 3 dB down beam width, with the accuracy of 0.17 degree. It was also estimated that the beam width did not change between before and after launch so much, but, SAR off nadir angle is 34.905 degree, which is 0.3 degree smaller than the original angle of SAR antenna deployment.
The bathymatric mapping is very important in various coastal studies such as biotrops inventory, animal or vegetation resources management, and pollution impact assessment. However, bathymetric surveying at shallow water areas by conventional shipboard sounding techniques is slow and dangerous and needs expensive devices such as a sonar and a ship with very shallow draught. Since the solar radiation penetrates to about 50 m in depth, particularly in clear water such as in coral reefs ecosystem or clear oceanic water, it is possible to extract the bathymertric information by using remote sensing data. This study describes the application of satellite remote sensing data for estimation and mapping of water depth. Remotely sensed bathymetry in the vicinity of the Kin Bay, Okinawa Main Island, Japan is performed and compared by using single, multi and bottom feature models. In these models, it was assumed that the ratio of bottom reflectance is the same for all points in the scene. For depth estimation, the constants values in the used models are computed by means of a multiple linear regression. Depth data ranged between 1 and 34 m at total of 377 and 374 points of Landsat-5 TM and SPOT-1 HRV imageries, respectively are extracted from a smooth sheet of survey charts received by JODC (Japan Oceanographic Data Center). The bottom feature model proposed in this paper yields a root mean square (r. m. s.) error of 3.07 and 2.67 m for TM and HRV, respectively. Based on this model, the 3-dimensional color imageries for TM and HRV sensors are derived and mapped by personal computer techniques.