2020-01-012024-05-15https://scholars.lib.ntu.edu.tw/handle/123456789/664571摘要：本研究以106年~108年的多尺度多元遙測資訊不確定性分析及時序性雷達影像水稻偵測技術發展為基礎，於109年度將配合宜蘭縣、雲林縣的水稻栽種時間，進行多尺度多元遙測資訊不確定性水稻偵測及指標優化。本年度除持續精進水稻分佈圖產製時效外，基於過去研究經驗，在同地區同期作但不同年份的水稻坵塊判釋中，於演算法所產出之指標權重，會因影像拍攝條件而有所差異。因此，109年將針對宜蘭縣、雲林縣水稻坵塊判釋作業所使用之影像指標進行貢獻度分析，以期找出於維持穩定判釋正確率高於90%之最少指標項目組合。藉由本研究的水稻偵測成果，將於稻作期間提供農糧署，藉以提升水稻分佈圖產製時效。<br> Abstract: This study will interpret rice fields distribution and index optimization in Yilan County and Yunlin County during two rice season in 2020 with the integrated results from multi-scale and multi-source remote sensing information analysis techniques developed since 2017. Based on the previous research experiences, due to weather condition were unstabile and unpredictable, the image interpretion indexes weights were different in the same method in each rice season. In order to stabilized overall interpretation accuracy over 90%, this research will analized the contribution of each interpretation indexes and composited into the most effective indexes data set for rice fields interpretation. With the interpret results of rice fields during two rice season in 2020, this study will assist Agriculture and Food Agency (AFA) to improve the producing efficiency of rice distribution map.多尺度多元遙測資訊不確定性分析水稻偵測Multi-Scale and Multi-Source Remote Sensing InformationUncertainty AnalysisRice Field Monitoring