A priori probabilities of landcover categories in the study area improve the landcover classification accuracy, although the probabilities are very difficult to estimate in advance before the analysis. Algorithms for the decomposition of mixels to pure landcover categories were developed to estimate landcover area ratios in mixels. If the study area is supposed to be a very large mixel which contains several landcover categories, some of the decomposition algorithms of mixed data can be applied to the centroid vector of the study area. The area ratios of landcovers in the study area are equal to the a priori probabilities of landcovers. The algorithm of maximum likelihood estimation was applied to estimate the a priori probabilities of landcovers in the study area in this research. As a result of this research, the estimation algorithm worked well and the a priori probabilities of landcovers in small study sites were estimated very well. Moreover, those estimated a priori probabilities of landcovers improved the accuracy of landcover classification in the study sites, comparing with the classification results of the maximum likelihood classifier, the Bayes' classifier with actual a priori probabilities and the italation of the maximum likelihood classifier.
The measurement method of a small-amplitude wave was proposed, in which a diffuse reflection spot was used as an index point. An equation to obtain the still water depth was introduced. This equation was confirmed experimentally by using a laser displacement sensor which is equivalent to a camera-index-point-film system in principle. To confirm the applicability of this method to the wave form measurement, the numerical simulation of the measurement by this method were carried out on sinusoidal waves and a composed wave. The results of these simulations show that the small-amplitude waves can be measured accurately enough when the water surface inclination is small enough.
In this study, the number of effective GLI bands and the combinations were selected to produce wetland vegetation map using GLI simulation data. In order to produce GLI simulation data, “rstar” was used as a atmosphere model and simulated up-ward radiance under the actual conditions of observing sun geometry and scan angles. To determine the GLI band combinations, training sample data were classified with maximum likelihood method by changing band combination of test data. Based on the classification accuracy of test data, optimal band combination was selected. In the case of 9 bands (GLI lkm, 250m channels) the first and second selected effective bands were : near infrared band (865nm) which is sensitive to the biomass, and short-wave infrared band (2210nm) which is sensitive to the water content. However, the case of using only GLI 1km channels (without short-wave infrared bands) selected bands were strongly effected by observing the sun geometry and scan angles.
An easy method for mosaicking JERS-1/SAR images was developed. According to the simulation for generating foreshortening distortion in side-lapped areas of two adjacent JERS-1/SAR images in range direction, it was proved that the differences of foreshortening distortion between the two images are decided uniquely only by surface elevation and they are not affected by the position in the side-lapped areas. It was also proved that the difference of the distortion changes in a linear relationship to the elevation value. An algorithm based on above characteristics of foreshortening distortion was tested using JERS-1/SAR images and digital elevation model with 250m grid from national digital land information and this algorithm worked well to create the mosaicked SAR image with no geometric discontinuity at the connection line of the two SAR images.
Camera calibration without control points or automated orientation are being received attention for the operatinalzation of digital photogrammetry. With this motive, the authors has been concentrating on decreasing control points by utilizing information such as distance which are included on the imagery or utilizing special equipment such as video theodolite. As an another approach for camera calibration without geodetic surveying, it is able to use dot pattern which are projected by dot matrix laser. This paper presents calibration method using dot matrix laser.
In the 8th year of the reign of Taisho (1919), with the arrival of the French tutorial unit, the Japanese Army started its aerial photographing. In the 12th year of Taisho (1923), the Shimoshizu military aviation school to which the French tutorial unit gave direct training, took aerial photographs of Tokyo immediately after the great Kanto earthquake. They played an important role in deciding the restoration plan. Triggered by this fact, 6 big cites including Nagoya, Kyoto and Osaka took the aerial photographs for the use of city planning, in the following 13th year of Taisho. The military aviation schools carried out the job in the guise of the military training. In the 8th year of Showa (1933), demands for the aerial photograph raised as application of the City Planning Act was extended to cover the new industrial cities. In the 12th year of Showa (1937) . the Japan-China war became intense and the cooperation from the military aviation schools went unattainable. The situation encouraged the birth of private companies undertaking the aerial photography. In the same year the Military Secret Protection Act was enacted and the aerial photographs were kept out of the public eyes as a part of military secrest. The city of Osaka has preserved the negatives taken in the years of 3rd and 6th of Showa in spite of the strict restriction and regulation of the Army.