水稻產量預估之航遙測技術研發

Rice Yield Prediction Using Remote Sensing Images

章國威(Kuo-Wei Chang) ； 申雍(Shen Yuan)

《航空太空及民航學刊.系列B》 43卷1期 (2011/04) Pp. 55-63

https://doi.org/10.6124/JAAA.201104_43(1).07

摘要

本研究擬結合已研發的水稻產量遙測模式、空載多譜影像、與福衛二號多譜影像，建立可預估田間水稻產量的遙測技術。研究中於農試所嘉義分所溪口農場設置水稻產量地真樣區約4.8公頃，種植梗稻和秈稻各一品種，並藉用不同氮肥用量等級和斷水乾旱處理形成產量差異，提供所需的產量地真資料。研究中除利用國科會精密儀器中心開發的VCDi和農航所航測飛機獲取地真樣區的遙測影像外，也使用福衛二號拍攝的遙測影像進行應用航遙測影像預估水稻產量之初步評估。研究結果指出，應用於水稻抽穗期間所獲取的航照和衛照遙測影像，預估收穫時的可能產量已具有可行性，若能再就遙測影像的大氣改正部分予以改進，將可進一步提高產量推估的準確性。本研究利用多譜遙測影像預估田間水稻產量的技術，將可以提高糧政單位對當期作水稻產量估測之準確度，有助於進行糧食供需調控的作業。

關鍵字

水稻Oryza sativa L. ； 特徵光譜 ； 產量推估 ； 遙測模式

並列摘要

This study intended to develop rice yield prediction techniques using air-borne and space-borne multispectral images. During this year, a 4.8-ha ground truth site for developing and testing the techniques has been established. Within the site, different rice yield scenarios were produced by using combinations of rice varieties, Japonica and Indica type, nitrogen rate and drought treatments. Remote sensing images collected by VCDi from PIDC and RSI from ROCSAT2 at booting stages were obtained and analyzed to predict the first and second crop rice yield of the testing site. The predictions were further verified by observed yield. Results indicated that techniques using remote sensing images at booting stage to predicted rice yield was in good progress. The prediction accuracy can be improved if better atmospheric correction method were used. It is expected that techniques to predict rice yield using remote sensing images can be developed. The accuracy and efficiency of current rice production estimation method can thus be upgraded.

並列關鍵字

Rice Oryza sativa L. ； Characteristic spectrum ； Yield estimation ； Remote sensing model