This paper describes an algorithm of a change detection on the land using the different features data. The conventional change detection method is roughly divided into two classes as follows: (1) multispectral image comparison (2) post-classification comparison The change detection has to use two temporal observed images or classified data such as Thematic Mapper images and land use maps. In general, the past data are the coded data such as land use maps or classified data, and the recent gathered data are the multispectral data. But there is the different sensor's specification data even if both of them are multispectral. Then an unadaption of specifications occurs in detecting changes between the different features data. These data specifications are not satisfied in conventional methods. It may be impossible to detect the land cover changes. In this paper is proposed a change detection method which is an effective comparison of thedifferent kind of specific characteristics data by incorporating the pseudo-spectral data. The pseudo-spectral data are given by adjusting the spectral characteristics between the compared data. The proposed change detection processing is constructed with pseudo-spectral calculation which transforms the coded data into the pseudo-spectral data and the features extraction in the spectral feature space. We show that the change detection accuracy of about 56% between categories of similar spectral features and about 90% between categories of dissimilar spectral features are obtained by the experiments using LANDSAT TM image.
A result of statistical approach to estimate both sea surface temperature (SST) and precipitable water from MOS-1/VTIR data collected over the sea around Japan is dercribed and discussed with emphasis on the usefulness of the data of the water vapor absorption channel. The various equations relating SST or precipitable water to the brightness temperature of VTIR channels along with the viewing angle, are examined in reference to the observed SST and precipitable water. The statistical analysis indecates that SST estimated from the brightness temperature at the two atmospheric window channels (ch. 3 & 4) shows a better result than that obtained from single channel data. It is found that the contribution of brightness temperature of the water vapor channel (ch. 2) to improvement of accuracy of SST is not large under cloudless condition. It is expected however that hardly recognizable thin cirrus is present the contribution of ch. 2 data will be significant. In the case of precipitable water estimation, however, contribution of ch. 2 brightness tempdrature to the improvement is significantly large showing an important role of ch. 2 brightness temperature is effective in estimation of the precipitable water.
An expression which estimates the total precipitable water from the infrared observed data of VTIR of MOS-1 is developed : W= α [1-I2/B2 (Ts)] +β [1-I4/B4 (Ts)] where I2 and I4 are respectively observed radiation value of the second band (67 μm band) and that of the fourth band (11.512.5 μm band) of VTIR, B2 (Ts) and B4 (Ts) are respectively the black body radiation values for a surface temperature Ts of the second and the fourth bands I2, I4 B2 (Ts) and B4 (Ts) are expressed by the observed brightness temperatures T2 and T4 of the second and the fourth bands of VTIR and surface temperatures Ts. And the surface temperature Ts can be also estimated from observed values of brightness temperature T3 of the third band and T4 of the fourth band of VTIR. The coefficients α and β in the above expression are given by the function of surface water vapor pressure Pv (0) : For Pv (0)≤2.91 mb α =0.3671, β = 26.95 For Pv (0)≥26.66 mb α =1.288, β= 32.38 For 2.91 < Pv (0) <26.66 mb α =0.3671+0.9209 [ (Pv(0) - 2.91)/19.75]4 β= 26.95 + 5.430 [ (Pv (0) - 2.91)/19.75]4 and for the case in which the proper surface vapor pressure is not available, the following expressions of α and β in terms of surface temperature Ts are given : For Ts ≤257.1°K α =0.3671, β= 26.95 For Ts ≥300.0°K α =1.288, β = 32.38 For 257.1 <Ts <300.0°K α =0.3671+0.9207 [(Ts-257.1)/42.9]6 β = 26.95+5.430 [(Ts-257.1)/42.9]6 The root mean square residuals between the precipitable waters from VTIR data and those from radiosonde data are 0.39 g/cm2/ (for α and β of functions of Pv (0)) and 0.51 g/cm2 (for α and β of functions of Ts)
Lake Chad is located in the Sahelian zone along southern boundary of Sahara desert. Expansion and reduction of water area were repeated several times since the prehistoric period. The authors found that the recent Lake Chad water area observed from satellite data is far smaller than that drawn on several maps. The authors investigate a change of Lake Chad water and vegetation area in recent ten and several years using LANDSAT and NOAA data. The mosaicked LANDSAT/MSS image of 1973 and seven NOAA/AVHRR data during the period 1982-1988 are obtained to analyze the distribution change of vegetation. Water and vegetation area analyzed from the VI (Vegetation Index) and NVI (Normalized Vegetation Index) obtained from each satellite data shows the decreasing tendency year by year. The close relationship between this change and the recent climatic data, such as annual precipitation, relative humidity and atmospheric temperature was recognized. This result may contribute to a monitoring of the present desertification in Africa. Concluding remarks; 1) The closed relationship was seen between the change of the area covered with water and/or vegetation and the annual precipitation change. 2) The area decrease corresponds to the decrease of relative humidity in the area and also the increase of the atmospheric temperature. 3) The decrement of the area reaches 40% during 10 years from 1973 to 1982. The area of water and vegetation in 1988 became 30% or less of that in 1973. 4) The recent Lake Chad is in a minimum size during the past 200 years.
The author intends to summarize here the present situation and future outlook of remote sensing data analysis with personal computer. With the rapid progress of the hardware and software and its price down, remote sensing data analysis system with personal computer has been practiced very often lately. High speed memory large enough to hold all information required, and easy operation language specially contribute to the construction of powerful data analysis system. The Remote Sensing Society of Japan has conducted the Training Course of Remote Sensing Data Analysis using personal computer three times in Tokyo and one time in Hiroshima. These activities are introduced first. Next, present situation of hardware and software for personal system design is explained. Finally, I mentioned the next generation of personal computer system for remote sensing analysis. It is predicted that the personal computer utilization for this field will become still more popularized. This special issue includes nine articles which introduce the present practice of remote sensing data analysis using personal computer in the following fields and subjects ; 1) Oceanography : NOAA/APT based monitoring of the Kuroshio. 2) Oceanography : Analysis of sea surface current and topography. 3) Meteorology : Typhoon and convective analysis. 4) Forestry : Forest-land type classification. 5) Agriculture : Air pollutant absorbed by agricultural land. 6) Environment : Environmental monitoring from satellite image. 7) Geography : A few examples of studies. 8) Construction : A practical application. 9) System : A system development with local area network.
We are now facing an exciting advancement in the geophysical fluid mechanicas of the Kuroshio, that has been accelerated by our real time satellite obtervation. Using a NOAA/APT receiving system based on a personal computer, we can apply routinely the flow visualization technique on the turbulence of the Kuroshio. The ready-made hard and softwares are used so effectively. The interpreters are supported by the data and knowledge that been built up in our indirect approach to the Kuroshio for 10 years. In this report, some NOAA images show a picture of the beautiful eddy structure dominating the ocean around Japan Islands. At a look, you can see that differing entirely from those of the conventional oceanography. Also, a hydraulic model test is introduced that shows the fluid mechanics of the eddy structure. The NOAA images, the picture of the eddy structure and the experimental results will stir up some mind of those who are interested in the oceanic turbulence in the sea.
NOAA AVHRR images of the Kuroshio off Honshuu were analyzed by a personal computer system. After the introduction of the system configuration, the method of estimating surface current velocities of the Kuroshio based on a single NOAA AVHRR image was developed and verified by the shipacquired surface current velocities during the period from 1986 to 1987. Although influenced seasonally and spatially, the method demonstrated the fairly good agreement between the estimated and the observed current velocities along the northern edge of the Kuroshio within the rms errors of about 0.5 knot for the maximum current speed of 3 to 4 knots and of about 14 degrees in current direction. Aiming for the integration of satellite remote sensing with geographic -information system, we also presented the integrated analysis of the AVHRR image with sea surface topography expressed as the geopotential anomaly data. By analyzing three dimensional data, that is, the sea surface temperature as two dimensions plus the sea surface topography as one dimension, we made it clear that the region of the highest temperature gradient corresponded well with the region of the highest surface gradient along the northern edge of the Kuroshio. In average off Toukai, one degree temperature difference equals about one geopotential anomaly difference, which turns out to be about 0.4 knot of geostrophic velocity. In order to strengthen both technologies of satellite remote sensing and geographic information system, more integration will be needed.
Authors have developed the software of the personal computer system, which has the functions of correction of distortion (geometric, radiometric), projection (polar) and display (cross section, histogram), to serve the system to general VISSR image data analysis. As some examples of application of the system, feature of typhoon and evaluation of convective cloud clasters were studied. The results are as follows: -the lowest pressure of typhoon occurred after maximum extention of central dence overcast (TBB lower than -70°C); -a feature of ratio of area occupied by every 5°C (TBB -40°C to -70°C) to the around typhoon center was changed after the direction of typhoon was suddenly turned northward; -correspondence between the highest reflectivity and the minimum TBB core was mostly good in the mature stage of a convective cloud cluster, but in the developing stage of the core the correspondence was poor; -the most ascending speed of a convective cloud top (325 m/minute) was observed in the developing stage. These results indicate that the system is effective for meteorological phenomena analysis.
This study shows a potential of personal computer to analyze the forest land with a satellite data. Although the SPOT data becomes near the airborne data in the spatial resolution, the geometrical distortion is large because of the slant viewing in some cases. Also, the terrain effect distortion cannot be deleted because that it is due to the geographical condition. To eliminate terrain effect distortion, the methods are considered; direct one and indirect one. The direct one is to calculate the geometrical condition of the sun and the slope rigorously. Indirect one is to use the ratio value of the adjacent channel data to avoid the effect. The improvement of categorization of forest type is attained higher in both corrected data case. To improve the categorization accuracy, SPOT HRV data was integrated with digital terrain model (DTM) for the forest land. The DTM are 20 meter mesh elevation data set obtained on 1 : 5.000 topographical map. The elevation is carefully picked up from the map within an accuracy of about one meter. Applying the DTM to the forest type analysis using the multi-spectral SPOT data, the categorization accuracy was improved about 20% above than that of the case with no application of it.
Recently, agricultural lands have been reconsidered from the view point of contribution for environmental preservation. Here, the function of these lands absorbing air pollutants is described as an example of such contribution. Concretely, two kinds of land use analysis using the agricultural census and digitalized air-photos for accounting areal existence of agricultural lands and their change are implemented for the purpose of estimating the absorbing quantity in case study areas. An adopted model is briefly explained at first, which estimates the quantity absorbing air pollutants of agricultural lands. Secondly, the method of applying the model to actual data set in case study areas, is mentioned, following the data processing flow. Finally, characteristics of the two data are discussed in the case of applying to estimation of air pollutants absorbed by agricultural lands. As conclusive remarks, the followings can be pointed out through case data processing using personal computing system. The agricultural census data will be helpful for a long-term trend analysis, an evaluation by a smallscaled administrative unit, and so on. On the other hand, remotely sensed image data will be adequate for grasping more detail change on both a time scale and an areal unit of lands.
Remotely sensed data, especially satellite images, have used for a wide variety of applications. The necessity of personal system for satellite image analysis has been recently recognized. Then the prototype system is developed for monitoring the environmental changes. Basic concepts of the system are as follows ; object oriented, friendly manmachine interface, cooperative use of existing software, and less logical/physical restrictions. Based on these concepts, the following features are implemented to the prototype system. 1) Handling capability of 3 bands with 1024 columns by 1024 lines images. Browsing is available by scrolling windows. 2) Image magnification by factors of 0.5 to 32. Multi-window is also available if the system has enough memory. 3) Pull-Down-Menu for function selections. Dialog-box is used to input characters or numbers, and select the function. 4) Full compatibility to the environment of the operating system such as Desk Accessary (resident programs) or Multitask. As the prototyping results, the future possibilities of a system based on a personal computer were demonstrated. This system is applied to the environmental monitoring. The temporal vegetation changes in Kanto Area are analyzed by two or three LANDSAT TM images, which were observed in different years. The loss of vegetations are especially clear in the Kawasaki and Tama. The basic characteristics of the thermal infrared images are examined in Miura peninsula. In this study, the implemented features of the prototype system proved to be very useful.
In this paper, we introduce an outline of our investigations of remote sensing which have been carried out recently in our department, some of which will be published soon. We treat two subjects as follows : 1) To clarify relationship between forest distribution and its controlling factors such as cumulative air-temperature, landform characteristics and snow cover duration using remote sensing images, DEMS (DTM) and climatological data. A technical problem to be solved was elimination of shadow in the imageries of mounstanous region by high precision DEMS. 2) To reveal time-series change of hydrologic environment in developing satellite city areas. The changes of the albedo and the expansion of impermeable surfaces, which are very important factors to control hydrologic cycle, were traced from multitemporal satellite imageries which were taken during past some ten years. Landsat imageries were overlaied with topographic maps and landuse maps to get hydrological parameters of varius ground surfaces of different land cover and different topographic conditions.
Recently, satellite remote sensing data have been felt to be familiar. But in the construction field, a utilization of remote sensing data is not a step of the practical application but the study yet. Accordingly, "Working group of the Civil Engineering Remote Sensing" has been established in the Remote Sensing Society of Japan, May 1989, for the purpose of widely applying remote sensing data in the construction field. This paper discribes an approach to a practical application of remote sensing data in the construction field by introducing a policy and course of "Working group meeting of the Civil engineering Remote Sensing". We have divided the operation of this working group into three routes and are studying them vigorously. Three routes are as follows : (1) Exercise of an image processing with a personal computer : For studying remote sensing and unifying a level of abilities among members, each member is exercising an image processing with a standard personal computer system, it is a system that is developed by remote sensing laboratory of Science University of Tokyo. (2) Collecting of literature : Members collect literature and data in connection with remote sensing, and arrange past applications of remote sensing to the construction field. (3) Image processing with a general purpose computer : For the porpose of the advancement of member's knowledge in connection with remote sensing, a specialist processes and analyzes remote sensing data with a general purpose computer and introduces the mothod of processing and their results.
Formerly Landsat MSS data was distributed on computer compatible magnetic tape (MT). This restricted image processing systems to high cost computer systems (mini-computers or larger computer systems) which included MT equipment. Since 1983 it has also been distributed on floppy disks opening the way for widespread image processing by low cost personal computers. This has usually been done by stand alone computers, however there are advantages in using a personal computer Local Area Network (LAN). All the personal computers on the network can share a huge block of MSS data, and can even access the same file at the same time. The authors have developed an image processing system for MSS data analysis by a personal computer LAN. This LAN system consists of a file server personal computer using a rewritable magneto-optical disk, which contains seven scenes -226 Mbytes-of MSS data, and ten personal computers, two of which have 768 Kbyte full color frame buffers for use interactive image processing.High processing speed was attained by C language programming. All menu selections are made with a mouse making it easy to use even for beginners. This system may be increased up to an optimum of four to five stations and still maintain a high level of system cost performance. This system may be highly usable for research in distant (up to 900 m) laboratories and other educational usage.