Remotely sensed data from satellite which are LANDSDAT Thematic Mapper and SPOT High Resolution Visible images can provide very useful information. However, the satellite image usually includes several kinds of distortions. To utilize them in high performance is necessary of the high quality image.
One of parameters for an image quality estimation is a spatial resolution for a sensor aperture characteristics.
This paper discusses the experiment to determine the actual spatial resoultion of an HRV image gathered by the linear array sensors carried on SPOT-1 satellite which was launched in 1986 and the anisotropic degradation of resolution due to the dwell time of sensor and the motion of satellite.
The spatial resolution are derived from a Line Spread Function which can be obtained by using a scene structure such as banks of field edges.
LSF (or Point Spread Function) is calculated by the differential of function which indicates a gray level of the edges.
In this paper, the spatial resolution is defined by the width at 50% fall point of PSF peak. Consequently, The numerical results of the spatial resolution are affected by the atmosheric effects, and the resolution in the along-track direction is worse 15.9% than that in the across-track direction. The linear array sensed image includes the degradations of image quality due to the sensor aperture characteristics and satellite motion. Therefore, The design of filters for image restortion are necessary to consider the direction.
by Y. Seto, and E. Nishijima

View full abstract

The infrared radiation in the spectrum of MOS-1/ VTIR band 2 can not penetrate through the atmosphere because it is located at the strong absorption range by water vapor. The data, however, reflects some condition of the top atmospheric layers. In order to know the degree of penetration of the light of VTIR band 2, we have done the evaluation of transmittance using the high mountain range data. A knowledge has been obtained that this light is transmitted from an altitude of at most 3000 meters to the satellite position. Based upon this fact, some vertical movements of water vapor are estimated using VTIR band 2 data. The results are compared with the direct measurement data obtained by radio-sonde and proved to be true. Third, we investigated some characteristics of the area where such water vapor depression occurs. The area is situated just north side along the jet axis. And the farther adjacent area northwestward from the water vapor depression area is covered with progressive high pressure. These phenomena are reasoned from the viewpoint of fluid dynamics in the earth rotational field.

View full abstract

A method to estimate semitransparent cirrus height together with temperature from the observed data of the water vapor absorption channel (Chan. 2 : 6.0-7.0um) and window channel (Chan. 3 : 10.5-11.5 um) of VTIR (Visible and Thermal IR Radiometer) onboard MOS (Marine Observation Satellite)-1 is developed.
The method is based on the radiative transfer model. In the derivation of the estimating formula following assumptions are made. (1) Scattering effect in the cirrus is neglegibly small in the spectra of these channels. (2) The temperature of cirrus and lower boundary (sea surface) is homogeneous in the area for computation. (3) The emissivity of the cirrus is equal in these spectra. (4) The temperature of the cirrus is equal to that of the ambient air.
Under the above assumptions, radiance of Channel 2 (R 2) can be expressed in terms of the radiance of Channel 3 (R 3) as follows.
R 2=aR 3+b, where a and b are determined through the least square method. Then the height of the cirrus can be decided through following procedures. (1) Compute radiances from the cirrus at the height of the initial guess for Channels 2 and 3. (2) Substitute the computed R 2 and R 3 into the above equation. If the equality doesn't hold adjust the initial guess height of the cirrus and repeat the procedure until the equality is reached. The height where the equality is satisfied is the cirrus height and the air temperature of that height is the cirrus temperature.
In the computation of the radiative transfer equation, the upper air data provided by the numerical weather prodiction group can be utilized.
Results of verifications indicates that there is a good correspondence between the height of cirrus estimated based on the proposed method and that obtained from the data of radiosoundings.

View full abstract

Remote sensing in agriculture has three applica-tion. One is the qualitative and quantitative analysis of lands features using the visible spectral and near infrared bands of Landsat MSS and TM data. This application includes harvest forecasting, disease information, and land cover classification. The second application is the gathering of information on soil water requirements by thermal infrared bands. The third application is assessment of the environment for the agriculture, e.g. estimation of evapotranspiration, surface temperature, and surface reflectivity.
In Japan, it is very difficult to make distinctions in land cover using satellite data because crop patterns are complex and the size of field is very small. Also the measurement of variable features in agriculture, e.g. harvest forecasting, is difficult because cloud cover renders satellite data useless. The stable features of land cover in agriculture, e.g. soil classification, were studied using satellite data. Recently, some environmental studies have been performed using satellite data, which is essential for the measurement of the environment over large areas.

View full abstract

Plants experience various kinds of stresses during the growth processes, including stresses caused by atmospheric factors, by rhizospheric factors, or by biotic stresses caused by other organisms. Besides, in recent years, it has often been reported that environmental toxic materials resulting from human activities inflict damage on crops and forests as in the case of acid rain.
However, there is no way to alleviate crop injury after symptoms appear on organs unless a large amount of agricultural chemicals is applied to prevent the spreading of the lesions. If farmers could detect various crop stresses, using sensors in the early stage of infection before visible symptoms appear, yield and quality of the agricultiral products could increase.
In this review article, I intended to describe the studies carried out to detect various crop stresses using ground-based radiometers, aerial photos or MSS at low altitude platforms.
Nutritient deficiency often appears on leaves, which can be detected by a spectroradiometer at the population level. Attempts of image analysis have also been made at the single leaf level. Water deficit can be clearly detected by using a thermoviewer which reflects the thermal changes caused by stomatal movement. For the same reason, disease infection can also be detected by the thermoviewer. Water deficit appears in the near- and mid-infrared wavelength reflectances in some crop species. For crop stress caused by soil-borne diseases, near- and mid-infrared bands can be used as they enable to detect the wilting of leaf.
Meteorological stresses sometimes do not appear early. Spikelet sterility is often caused by low temperature during the flowering stage of rice plant in the northern part of Japan, while the color of the ears remains green at harvest time in autumn. Rice damage caused by flooding is highly correlated with the turbidity of water. Since the effects of these meteorological phenomena can be observed by satellite sensors, maps of damage to crop yield can be displayed using techniques of image analysis.
Stress sensors have been mainly developed in the field of physics, representated by magnetic wavelength reflectance. However, in the future, chemical sensors will become important to monitor the crop growth environment in real time. Furthermore, the use of bio-sensors will become more popular for the detection of bio-information due to their high sensitivity and selectiveness.

View full abstract

The air temperature near the ground is very important to the growth of the rice plant becuase of its tropical origin. Therefore, information of the air temperature is very important for the cultivation of the rice plant. The purpose of the present study is to estimate the 1.5-m height air temperature in region where there is no meteorological observation site. The temperature estimate will be determined using satellite thermal IR data.
First, field measurements were made to determine the correlation between the air temperature and the surface temperature. The air temperature closely compared with the surface temperature, and the correlation coefficient for all the data was over 0.90. Further analysis was done to relate the air temperature to other meteorological parameters ; net radiation, humidity and wind speed. However, the best correlation coefficient was obtained using surface temperature. This suggests that the air temperature can be accurately estimated by the surface temperature from the satellite IR data.
Secondly, the surface temperature from the GMS IR data derived using equations (1) and (2) was compared with the AMeDAS air temperature in Hokkaido. The maximum correlation coefficient between the AMeDAS air temperature and the surface temperature determined using the GMS IR
data was 0.90, the minimum was 0.30 and the mean was 0.65±0.15.
Thirdly, the air temperature distributions in Hokkaido were estimated using the regression equation which relates the AMeDAS air temperature and the surface temperature determined using the GMS IR data (Fig. 3). The air temperature distributions in Ningxia, China were also estimated by using the GMS IR data.
Fourthly, the hourly mean deviation was calculated for the regression line obtained using the surface temperature from the GMS IR data and the AMeDAS air temperature in order to investigate the characteristic of the ground surface condition for each of the AMeDAS sites. The AMeDAS sites were classified according to their patterns for the hourly mean deviation. Using the results of this classification, the temperature distributions of the five areas in Hokkaido were estimated. The accuracy of the estimation using this method was 1.0±0.1 (^°C).

View full abstract

The aim of this paper is to present a making method and some applications of the soil organic matter content (hereinafter referred to as O.M.) map of Obihiro Area, Hokkaido, made by using Landsat TM data. The O.M. of each site is derived from the correlation between measured O.M. in the field and CCT value of TM band 3. The O.M. of vegetationcovered cropland is estimated from CCT value of band 3 on the soil line transformed by the soil index method (Fukuhara et al., 1979). The map consists of eighteen sheets and covers whole cropland of Obihiro Area (1, 083, 000 ha). The map is characterized by four points as follows; (i) grid cell size is 25 m×25 m, (ii) the O.M. is devided into seven classes (0-2%, 2-5%, 5-8%, 8-10%, 10-12%, 12-15% and 15-20%) and shown on 1 : 50, 000 to pographical maps in seven colors (red, brown, yellow, light green, cyan and violet), (iii) users can make a desired scaled map since this map is also provided in floppy disks, and (iv) users can see laying it on top of our map since the soil map is appended as the overlay sheet. Users can know the O.M. of the desired point from this large-scaled map in detail. When our map is used together with the soil map, it is of great use for some purposes such as scheduling irrigation, improving draining, improving soil and managing crop.

View full abstract

The relationship between ground surface temperatures on clear nights in the late frost season and altitudes was analyzed. Data for ground surface temperatures were taken from NOAA/AVHRR infrared channel data (NOAA data) and those for altitudes from the Digital National Land Information of Japan (DNLI). The relationship showed that the hypothetical temperature of a ground surface (θho), if the ground surface would be lowered adiabatically from the altitude of h to 0 m, is calculated from the following equation : θho=θ+A·h/ 100, where θ is the ground surface temperature in ^°C, h the altitude in m, A the regression coefficient or the lapse rate of ground surface temperature in ^°C/ 100 m. The hypothetical temperature is defined as the potential temperature of ground surface.
The potential temperature (θho) of a pixel was calculated by the ground surface temperature from NOAA data, the altitude from DNLI, and the regression coefficient (A). Three color images of potential temperature on three clear nights in April to May were drawn. The images showed that the relatively warmer pixels locate in the terrace terrain, the relatively colder pixels in the basin terrain, and that the intermediate temperatures are in the hill slope and the flat terrains. The similarity of the spatial patern of three images implies that the potential temperature of ground surface is an usefull index in assessing the temperature environment of the ground surfaces on clear nights in the late frost season, and for detecting the places which may be susceptible to lower temperatures and/or late frost damage, or which may be of advantage to crop production.

View full abstract

A number of investigators have conducted experiments that used radiative surface temperatures to evaluate the evapotranspiration (ET) on spatial scales from a field plot (using hand-held radiation thermometers) to a regional size (using certain satellite sensors). The purpose of this paper is to review those studies and to present some of the author's recent results, employing a simple estimation model of ET above vegetation.
Studies concerning estimation of ET using remotely sensed thermal measurments were classified into two categories according to platforms of measurements. The first group includes studies which used ground-based measurements to develop models for estimating ET or monitoring the moisture condition of plants or soil. The other group includes studies which used data obtained by air- or spacecraft to estimate soil moisture or ET on a regional scale.
The authors tested the modified equilibrium evaporation model to estimate ET above crops and forests. This model uses surface temperatures of plant canopies to evaluate equilibrium evaporation rather than using air temperatures that have been a convention. The ET estimation is calculated by multiplying the surface-dependent parameter by the equilibrium evaporation, which is a function of surface temperature, net radiation and soil heat flux. Field experiments were conducted to obtain model parameters for a wheat field, a rice paddy and a broad-leaf forest on fine days under drought-less condition. Equilibrium evaporation values cal-culated using canopy surface temperatures were compared with measured ET values using the energy budget with Bowen ratio method. The results indicated model parameters of 1.03 and 1.06 for both kinds of fields and for the broad-leaf forest, respectively.
This model was applied to satellite data (NOAA/ AVHRR data) for estimating ET of the broad-leaf forest under droughtless conditions. The estimated values agreed well with measured values, and thus the model was proved effective.

View full abstract

Evapotranspiration (ET) is one of the major elements in the water circulating process on the earth. From the view point of water resources or water utilization, the distribution of ET over wide area in a long term, like monthly, should be considered in connection with the regional characteristics of topography or landuse. Remote sensing technique is a probable means to estimate the distribution of ET over wide area, however this technique has some difficulties to measure ET directly for a long duration.
Here the model which adopts an indirect method to estimate ET has been developed by using Landsat MSS, digitized elevation and observed meteorologi-cal data near the ground. In order to examine the effects of topography or landuse conditions on the amount of ET, some calculating experiments have been executed by this model, which is able to estimate ET under arbitrary presumed conditions. The results of calculation are summarized in the following.
With regard to the topographic conditions, the amount of ET is scarecely affected by the variance of elevation unless the area has dominant slope orientation. On the other hand the value of elevation averaged in the area is found to be a significant indicator of topographic conditions on ET.
Landuse conditions are represented by reducing to two empirical parameters in the model. The regional averaged values of these parameters might be choosen as representative indicators of landuse in wide area relating to the environmental condition on ET. This result suggests that there is a possibil-lity to trace the landuse change in the past and to assess the conversion in the future from the hydrological point of view by adopting remote sens-ing technique.

View full abstract

Land improvement project is supported technically by "Nogyo-doboku". Nogyo-doboku is the name of irrigation, drainage and reclamation engineering which has made progress in Japan as a system of technology for agricultural infrastructure development. In these days, a subject of Nogyo-doboku spread in the field of life environment and even in that of natural environment. Nogyo-doboku is developing into new technology for creation of beautiful country and rural districts where people can live in abundance.
Therefore. remote sensing will be useful in the field of Nogyo-doboku, and it will be available at the time of operating of land improvement project. Especially, it will be of use to investigation for planning of project or monitoring for management of rural district after land improvement project.
We wish and expect the advancement of remote sensing and image processing technique to extract various information on the earth. In this report, we made mention of the application of satellite remote sensing data for investigation of crop damage. The distribution map of rice damage due to cold weather in Ishikari plain was made by using Landsat MSS data. And the map of beet damage due to dryweather in Shari-Abashiri district was made, too. Information extracted from maps like these wil be useful not only for agricultural management but also for planning of land improvement. Next, the accuracy of land use classification in paddy field area was mentioned. As a result of analysis, it became clear that almost all error in classification occured at boundary of land use. Furthermore, monitoring of natural environment around Kushiro-shitzugen using remote sensing was mentioned. This area is peat land area and a view is on a grand scale. We can see many valuable aninals and plants in this area. The changed place of land use or natural vegetation cover during the past seven years was discriminated by overlapping analysis of Landsat data and MOS-1 data which covered Kushiro-shitzugen.

View full abstract

VISSR (Visible and Infrared Spin Scan Radiometer) image data observed by GMS (Geostationary Meteorological Satellite) have been stored since March 1981 in the MSC (Meteorological Satellite Center). Since storage media and data formats used in the MSC are designed for a large-scale computer system, those data are yet in limited use.
For easy processing of VISSR image data on personal computers, a procedure to convert image data from MT (Magnetic Tape) or CT (Cartridge Tape) in the MSC into floppy disk by resampling has been developed since 1988 by the author. Following problems, however, are arisen in resampling from original VISSR image data.
1. Ratio of pixel number density on a floppy disk to that of original VISSR changes with location on the earth.
2. Observed scan lines are not usually parallel to the latitudinal line and therefore some information on each scan line is more or less obscured. Users can not exactly extract observed time and channel name (on the visible data) of a pixel.
3. A usable image on a personal computer is necessarily limited in region and resolution, resulted from the preceding resampling.
In order to solve these problems an image mapping technique with a new format of the image data for personal computers is developed. This mapping is derived without losing accuracy by a simplified mapping method currently used for a large-scale computer in the MSC.
These image data on a-floppy disk, though the size to be treated is necessarily small, make it possible for personal computer users to deal with VISSR image data in original resolution both on space and temperature (or reflective intensity). Personal computer user then will be able to analyze in more detail the field of the sea surface temperature and/or cloud distribution of various meteorological phenomena such as meso-scale disturbance.
Some example of application of technique de-scribed are also shown in this paper.

View full abstract