2014-01-012024-05-17https://scholars.lib.ntu.edu.tw/handle/123456789/684272摘要：依據前期計畫「航遙測技術輔助休耕田搶種之監測調查」執行成果中，發現若僅以每期稻作之插秧期與分蘗盛期影像資料進行休耕田違規搶種判釋，於實驗區中雖可達成88.44之整體正確性，但對於違規搶種的物種分類上仍有待持續研究。本年度計畫除延續前期計畫，持續辦理航遙測技術輔助休耕搶種之監測調查，並嘗試藉由農航所建置之機載合成孔徑雷達系統所拍攝之雷達影像資料，與光學遙測影像資料進行整合，藉由雷達影像可穿透雲層之特性，輔助光學影像資料判釋上所不足之處。 本年度計畫在光學遙測衛星影像資料方面，將由原本每期稻作兩組影像增加為四組影像，藉以掌握更細緻的申報休耕坵塊範圍之地上物光譜變化特徵；在偵測對象方面，除明顯偵測違規之申報休耕坵塊外，針對第二期稻作所發現搶作之大蒜作物，本計畫結合大蒜種植區域之監測與複查機制，除比對鄉公所調查資料與遙測結合地面調查之監測成果，建立複查機制外，並研究搶作大蒜之坵塊光譜特徵；在智慧型農田坵塊單元萃取模組部分，以像元式及區塊式資料轉換特徵向量型態分析，持續進行提高坵塊精準性之研究。藉由本年度計畫預定執行之光學遙測影像稻作與休耕作物判釋、整合光學與雷達影像進行稻作及休耕作物判釋可行性評估、搶作大蒜光譜特徵及偵測模式研究、大蒜種植區域之監測與複查機制、智慧型農田坵塊單元萃取，除可持續發展遙測作物判釋技術，並可結合作物複查機制，協助農政單位更有效且可靠的掌握農地耕作現況.<br> Abstract: According to the results of previous year project, we found that if only use transplantation stage and tillering stage images to interpret fallow fields are legal or not, although the overall accuracy is 88.44%, the classification of illegal crop still have to be improved. In this year,base on the previous year results, we will continue to apply remote sensing techniques to monitoring illegal fallow fields and try to combine airborne SAR image data, generated by Aerial Survey Office, and optical remote sensing data to interpret rice crop and fallow field crops. The improvements of this year, in optical satellite image part, we will use 4 pair of image data from rice transplantation to harvesting to understand more detail about the spectrum change of crops in fallow fields. In monitoring target part, except obviously illegal fallow fields, we will focus upon illegal fallow fields with garlic. This project will integrate garlic monitoring and reexamine mechanism, compare garlic crop data between generated by township office and monitoring by remote sensing interpretation with ground survey to reexamine the data accuracy and analyze the spectrum feature of illegal fallow fields with garlic. In smart farmland unit detection model part, we will apply pixelbased and regional-based data transfer feature vector analysis to improve the accuracy of farmland unit detection. With integrating optical and SAR image data in rice crop and fallow field crops interpretation, illegal fallow fields with garlic spectrum feature analysis and interpretation, establish garlic crop monitoring and reexamine mechanism, smart farmland unit detection model in this project, we can continuing developing remote sensing interpretation techniques of crops and integrating crops reexamine mechanism to help Agriculture and Food Agency (AFA) monitor farmland more efficient and reliable.影像分割影像分類影像單元化知識庫管理農田耕作狀態轉移偵測休耕田遙測技術機載合成孔徑雷達系統Image SegmentationImage ClassificationImage UnitKnowlage Databased ManagementFarmland cultivation state transition detectionFallow