Detection of land cover change by time-series satellite data for the extraction of environmentally changed area in the whole Asia

利用1981-1999年的NOAA AVHRR数据对整个亚洲的土地覆盖变化进行了研究。通过本研究，取得了以下成果。(1)提出了一种去除AVHRR NDVI数据云层影响的新方法。该方法称为时间窗操作(Two)方法。利用该方法得到了自然平滑的季节NDVI曲线。(2)阐明了太阳天顶角对AVHRR NDVI的影响。研究发现，当太阳天顶角超过60度时，很难对其进行校正。(3)求出了8 km AVHRR数据的最小可探测地面面积。(4)利用1981~1999年的AVHRR NDVI和地表温度数据，提取了全球范围内植被增加/减少的区域。为了进一步研究，为了绘制更可靠的土地覆盖变化信息，强烈建议建立地面实况数据库。

The study to detect land cover change of the whole Asia was carried out using NOAA AVHRR data from 1981 to 1999. The following achievements were obtained through this study. (1) A new preprocessing method to remove cloud effect included in AVHRR NDVI data was developed. The method is called Temporal Window Operation (TWO) method. By this method, natural smooth curve of seasonal NDVI is obtained. (2) The effect of solar zenith angle to AVHRR NDVI was clarified. It was found that it is difficult to correct the effect of solar zenith angle when it exceeds 60 degrees. (3) The minimum detectable ground area by 8 km AVHRR data was found. (4) Areas with vegetation increase/decrease were extracted for global area using AVHRR NDVI and surface temperature data from 1981 to 1999.For further study, in order to map more reliable information on land cover change, the development of ground truth database is strongly recommended.