水稻植被反射光譜之特徵及變異

Abstract

本硏究於2000-2002年（計有四期稻作）在行政院農業委員會農業試驗所農場進行田間試驗，量測近地面水稻全生育期之植被高解析反射光譜（350-2400 nm)，以探討不同期作及年度間反射光譜之特徵及變異，提供光譜地眞資料作爲精準管理等用途。爲顯示水稻全生育期之反射光譜變化輪廓，乃將高解析植被光譜以三維動態曲線展現’發現其與 一般綠色植被光譜曲線具有相似性。將光譜區分爲紫外光（UV，350-400 nm)、可見光（Vis， 400-740 nm)、近紅外光(Nir，740-1300 nm)、短波紅外光(Swir，1300-1800 nm)及中紅外光(Mir, 1800-2400 nm)等五寬頻波段比較，結果以近紅外光波段在全生育期的變異最大，紫外光波段最小。續將可見光進一步細分爲紅光(640-740 nm)、橘光(585-640 nm)、黃光(560-585 nm)、綠光(490-560 nm)、藍光(425-490 nm)及紫光(400-425 nm)等波帶，發現以紅光帶變異較高，紫光帶變異較低。從光譜曲線篩檢動態變異的綠光波帶峰點(RGREEN)、紅光波帶谷點(RRED)及近紅外光波段頂點(RNIR)等三項光譜特徵反射比，由其等在生育期間的變化可以反應出在不同期作及年度的差異。此三項光譜特徵之間的比値、RNIR-RRED、標準差植被指數（NDVI)及紅光臨界區間與斜率等，在生育期間均呈現二次曲線分佈，顯示其等與生長之間具有互動關係，具有作爲監測水稻生長狀態的應用潛力。 Field experiments were conducted in the experimental field of Taiwan Agricultural Research Institute to obtain ground-based remotely sensed rice canopy reflectance spectra (350-2400 nm) of high resolution in the cropping seasons in 2000-2002. Characteristics of spectra were investigated and variations in different cropping seasons and years were studied to provide the ground truth for further precision management applications. The three dimensional graphs were plotted to show changes of spectral curves during the growing periods and between seasons. As indicated, the reflectance spectra of rice canopy were similar to those from other green plants. By dividing the spectrum into 5 broad bands, the greatest variation was observed in the near-infrared (Nir, 740-1300 nm) and the smallest variation was in the ultraviolet (UV, 350-400 nm). By dividing the visible light into red band (Red, (>40-740 mu)- orange band (Orange, 585-640 nm)，yellow band (Yellow, 560-585 nm), green band (Green, 490-560 nm)，blue band (Blue, 425-490 nm) and violet band (Violet, 400-425 nm), it showed that red band had the highest variation and violet band had the lowest variation. Reflectance of the spectral characteristics at green band maximum (RGREEN), red hand minimum (RRED), and infrared peak (RNIR) showed a curvilinear variation during the growth and a difference between seasons. Ratios from these three characteristics, NDVI, RMK-RRI D, red edge slope and red edge span were also changed curvilinearly during rice growth. Variations of these parameters have their physico-physiological meanings and arc related to plant growth and environmental influences. The connections to the morphological progresses of rice plants make these parameters potential in applications for growth status monitoring.