NDVI analysis in Korea with NOAA AVHRR images

Abstract

In North Korea, the shortage of food has been a serious problem for the past several years. However, the actual condition of present North Korea cannot be investigated there. Then, it analyzes about the food situation of North Korea from a viewpoint of vegetation using the remote sensing technology using the satellite. In Tohoku University, NOAA AVHRR images are accumulated over the past ten years or more. It can study for the difference between present and the past in analyzing vegetation over a long term. We analyze vegetation in Korea using NOAA-11,14 and 16 from 1989 to 2002. As a satellite is employed over a long term, a sensor will deteriorate. So, it is necessary to calibrate degradation of a sensor. And also the observation range of wavelength is different for some NOAA satellites, we calibrate the data to keep a reliability. I. INTRODUCTION The shortage of food has been a serious problem for the past several years in North Korea. However, the actual condition of present North Korea cannot be investigated there. If we can analyze about the food situation of North Korea from a viewpoint of vegetation usign the remote sensing technology using the satellite, it is very useful data for food support. In these years, FAO(Food and Agriculture Organization of the United Nations) started observation a vegetation around Korea using SPOT data. It is possible to analyze to go back more than ten years when come being able to analyze vegetation using NOAA AVHRR images in this study. And so, this study observated the transition of NDVI(2) in Korea and considered to relation of meteorologial phenomena and transition of NDVI. II. GEOMETRIC CORRECTION The NOAA AVHRR images is not good accuracy of loca- tion of observation. The analyze for long term using NOAA AVHRR images is need to geometric correction. We developed fast and high accuracy geometric correction method with NOAA AVHRR. The geometric Correction has two approach, correction of satellite position and correction of image position. The first approach is need mightly calculation time. This study only need correct data around Korea. Moreover, it use mightly number of data about ten years or more. For that purpose, we selected the second approach.