Mitsuru Miyazato

Use of rice response characteristics in area estimation by LANDSAT/TM and MOS-1 satellites data

Abstract Satellite image interpretation for LANDSAT/TM and MOS-1 data was conducted for identification and area estimation of rice fields for Hishikari Town in Kyushu Island near Kagoshima City, Japan. The results were confirmed by video-recorded aerial photographs and by ground observations. The rice field identification, which was carried out by computer-implemented MDM classification using LODIA software, created some confusion between different test sites. However, the digitally processed image of the LANDSAT/TM 4 and 5 band difference index, discriminated the rice fields very well. The estimated rice areas were observed to be fairly close to the statistical reports made by the town office. The present study shows the potential of image processing for the identification and area estimation by rice response characteristics.

Simulation of Survival Survival Curves in Several Molds Irradiated with Gamma and Ultraviolet Rays.

It is expected that gamma (γ) and ultraviolet (UV) rays are applicable to sterilize agricultural products for reducing storage losses. In practical application, D10-and D99.9-value in γ and UV irradiation have been used as sterilization indices. However, our preliminary results showed that these survival curves were exponential having a shoulder at a beginning in many strains of molds. This suggests the difficulty in assessing a sterilization level accurately using these indices.The survival curves of several molds infecting agricultural products for γ irradiation were simulated with a multi hit theory in molecular radiation biology. On the other hand, the effect of UV irradiation on the survival curves was simulated by multi hit theory, taking into account the decrease in UV-intensity in liquid suspensions due to Lamberts law. It was found that the models based on a multi hit theory well simulated the effects of γ and UV irradiation. Among molds tried in this study, the four hits model best fit to the observed data in cases of Phycomycetes.These results demonstrated that D10-and L-values, instead of conventional indices based on D10-and D99.9-values, should be used to assess the sterilization level accurately. To obtain more reliable measurement, a target area and a hit number, a parameter of penetration in UV-sterilization should be taken into considerations.