The objectives of the study are to develop a system to measure the data of wind flow in both the horizontal and vertical directions utilizing stereo photogrammetry, and to establish the methodology for three dimensional measurement on site of wind flow near the ground surface.
This system uses two cigar-shaped kite balloons equipped with a 35mm still camera or four 35mm still cameras set on the ground, which stereophotograph a third balloon released as a tracer, thereby three-dimensionally analysing the third balloon's path of particle.
As the results of the experimental study on the campus of Chiba Institute of Technology and Kamogawa, Chiba, paths of particles of wind flowing over buildings and mountainous regions were successfully observed.

抄録全体を表示

Aerial photogrammetry is powerful technique to monitor landslide behavior effectively and safely. Using three dimensional displacement (3DD) vectors of ground points, which are detected by photographs taken at different time flights on a landsliding area, velocity and direction of ground movement can be visualized on a map or projection plane. Moreover, we can presume geometry of underground slide surface by analysis of the 3DD vectors. This paper treats a landslide that had abruptly occurred at Hirose in Nagano city. Its characteristic was small-area, rapid-growing, short-time movement. Survey-analytic system applied to this landslide demonstrates validity of the aerial photogrammtry.

抄録全体を表示

Abstract : 3-D measurement of rice canopy is effective for growth analysis, photosynthesis evaluation, and reflectance modeling. But, digital measurement has been hardly carried out because 3-D measurement of plant canopy is difficult.
In this paper, past studies on stereo matching are reviewed and it is shown that not only information of similarity but also that of continuity is necessary in order to match each point of rice vertical images. Therefore, a new matching algorithm for rice plant images is proposed. The algorithm is based on recognition of the top of a leaf and a following outline of the leaf. Rice 3-D shape was measured presicely with this algorithm.

抄録全体を表示

Canopy reflectance models were built in order to relate remotely sensed data to canopy physical state (canopy structure, leaf reflectance) . Most of these models are based on canopy parameters such as vertical leaf-area density, and leaf normal distribution function. Using these models, canopy physical state is inverted from remotely sensed data. But, theoretical consideration on the accuracy is not enough. This is because the precise canopy 3-D structure is not dealt with.
Therefore, rice 3-D structure is measured with the stereo photogrammetry and the leaf edge matching method, which was proposed in our last paper. Then, the 3-D structure conserning to photon transport in canopy is analysed with growth.

抄録全体を表示

This study is aimed at the analysis of the earth's surface character and data processing methods on the extraction of Lineaments from TM VNIR data taken by Landsat 5. It is generally understood that to interpret the geological structure of the target area is the basic issue for data analysis to explore the oil and gas using remotely sensed data. Earth's surface topography mostly reflects the structural motion of the crust and indicating the geology. Lineament patterns are the most significant features as well as geological information which normaly behave as the index to indicate the historical results of structural motion. Image enhancement of satellite remote sensing data is one of the most effective method to discriminate the lineaments which indicate the surface topographic features reflecting sub-surface structure.
In this paper, three Visual Near Infrared region of Landsat TM data with different types of surface character, vegetated mountainous, vegetated flat and mountainous rock exposures, were chosen and three types of digital filter were applied in order to enhance the images to extract and analyze the lineaments from each images. Types of the filters were selected by the points of enhancement and preservation of the edges which contain in the imagery. Lineaments were extracted using 12 processed images and lineament patterns were quantitatively evaluated. As the results, the number of total cumulative length of edge preserved images of median filter are in maximum. Median filter images preserve the lineament pattern which recognized in non-enhanced images, but high frequency component enhancement images were not preserved. This results suggest that using median filter is one of the effective method to interpret remote sensing data.

抄録全体を表示

There are many different types of cheap disposable cameras, such as e.g. the 35mm type or the panorama variant, which are becoming more and more popular. These cameras can also be used for 3D measurement and as that use is becoming popular, there is a need to assess the accuracy of such measurements.
This paper presents a method for calibration of disposable cameras for 3D measurements, by comparing the measured points with a number of check points. The results show that the X and Y accuracy is roughly similar to that of a single lens camera, indicating that disposable camera may become convenient instrument for 2D measurements.

抄録全体を表示

In global environmental studies, GIS-based computer environment (global GIS) is keenly demanded to support handling a huge amount of data, integration of a wide variety of global data and spatio-temporal analysis. A spherical tessellation scheme with which a spherical surface is tessellated into regular grid forms one of the important basis of a global GIS. Conventionally, many of global datasets are developed based on equal division of longitude-latitude coordinate system. However, area distortions and radial distortions especially in high-latitude areas may cause inconvenience and inefficiency in area calculations and neighborhood analysis. In this study a quantitative comparison of spherical tessellation schemes for global GIS is made with respect to a set of criteria to clarify their strengths and limitations. Through the comparison, North-Up ZOT projection proposed by the authors is found relatively suitable for global GIS especially when existing GIS software are used for global GIS.

抄録全体を表示

NOAA-9からNOAA-11への移行期を含む1986年～1989年の48ヶ月間のmonthly GVIデータに対し主成分分析を行い, 得られた主成分の性質を調べた。その結果, 以下のことがわかった。第6主成分にNOAA衛星移行の影響が現れている。第1主成分はNDVI値の総和に相当し, 第2, 第3主成分は年間1周期変動を表す。第4, 第5主成分は年間2周期変動を表す。分散共分散行列による方法と相関行列による方法とを比較した場合, 年間周期変動を示す第2～第5主成分ではその違いは少なかったが, 第1, 第6主成分ではより大きな違いが見られた。

抄録全体を表示

Video image sequences often give important information about the dynamics of human motion in the field of sports, training or rehabilitation. In understanding of the dynamics of human motion from general TV or video image sequences, there are two complicated subjects. One is image processing, for example automated recognition of features on the human body, such as e.g. the head, elbows or knees. The second is how to estimate the photogrammetric camera orientation parameters on a moving camera.
This paper describes the development of a video theodolite system, consisting of a CCD camera, a theodolite and a video recorder. The system makes it possible to record image data with a moving camera and simultaneously determine the camera rotation parameters in real time.
The authors first analyzed the dynamics of the sprinter Carl Lewis and that of boat rowing by using TV images. In both these cases however, the camera orientation parameters could be determined by utilizing fixed information in the images such as the goal line and course lines in the first case, and the buoys marking the rowing course line in the second case. But often, there exist no such fixed information in the images why the authors concentrated on developing a system where the camera orientation parameters could be determined in real time while recording a moving object.
In order to measure the rotation parameters, a CCD camera was mounted on top of a theodolite.ωand κwere defined as the horizontal and vertical angles respectively, φalways being 0 degrees as the theodolite was levelled. The current values of the parameters were then superimposed continuously on the image frames and thus recorded as a part of the image data.

抄録全体を表示