Publications-Theses
Article View/Open
Publication Export
-
Google ScholarTM
NCCU Library
Citation Infomation
Related Publications in TAIR
題名 農作物空拍影像種類辨識與無人機農檢任務應用效能評估
Assessment of Crop Aerial Image Classification and Performance Evaluation of Drone-Based Agricultural Inspection Tasks作者 蔡孟宗
Tsai, Meng-Tsung貢獻者 劉吉軒<br>彭彥璁
蔡孟宗
Tsai, Meng-Tsung關鍵詞 無人機
人工智慧
深度學習
農作物辨識
物件分類
語意分割
UAV
artificial intelligence
deep learning
crop identification
object classification
semantic segmentation日期 2025 上傳時間 3-Mar-2025 14:29:16 (UTC+8) 摘要 本研究旨在建立本研究旨在將無人機與AI模型整合,進行農作物種類辨別以及農田面積估算。使用無人機收集欲辨識之農作物照片及高空農田照片為資料來源。以收集之農作物照片,對物件分類模型(Zhao等人,2017)進行微調,此模型用以農作物種類辨識之用。以收集之高空農田影像,對語意分割(Wang等人,2018)模型進行微調,此模型用於控制無人機飛行高度,飛行高度需達到可辨識所指定之完整農田面積。 系統成效評估是基於以下原則,準確性、效率性、安全性。首先,系統需具備高準確度,以確保農作物種類辨識及農田面積估算的精確性。其次,系統需具備高效性,以快速完成巡查工作,減少人工干預。最後,系統需確保巡查人員的安全,降低人員出行的風險。 本研究的主要貢獻在於提升農田作物巡檢效率,使用無人機替代人工巡檢行為,從而降低人力成本與時間以及巡檢之風險。透過微調的物件分類模型,進行農作物辨識,為農田作物種類提供更可靠的數據支持以及記錄相關影像作為日後查驗。同時,使用語意分割模型的應用使農田面積估算,研究使用語意分割模型進行無人機飛行高度控制。此外,本研究優化了農業補助申請流程,透過自動化檢查與數據記錄,提高巡檢工作的效率與準確性,減少人工作業的誤差。無人機與 AI 技術的整合為農業應用提供了實際資料佐證,推動精準農業的發展,同時降低巡查人員的安全風險,並減少農業活動對環境的影響。最終,本研究為建立可持續的農業巡檢系統奠定基礎,助力實現智慧農業與可持續發展的長遠目標。
The purpose of this study is to integrate drones with AI models for crop type identification and farmland area estimation. Drones will be used to collect photos of the crops to be identified and aerial images of the farmland as data sources. The collected crop photos will be used to fine-tune the object classification model, which is intended for crop type identification. The collected aerial farmland images will be used to fine-tune the semantic analysis model, which is used to control the flight altitude of the drone, ensuring the altitude is sufficient to recognize the specified complete farmland area. The effectiveness evaluation of the system is based on the principles of accuracy, efficiency, and safety. First, the system must have high accuracy to ensure precise crop type identification and farmland area estimation. Second, the system must be efficient to complete inspection tasks quickly and reduce manual intervention. Lastly, the system must ensure the safety of inspection personnel, reducing the risks associated with their fieldwork. This study enhances the efficiency of crop inspection in agricultural fields by utilizing unmanned aerial vehicles (UAVs) to replace manual inspections, thereby reducing labor costs, time consumption, and operational risks. Through the fine-tuning of an object classification model, the system enables accurate crop identification, providing reliable data support for crop classification while recording relevant images for future verification. Additionally, the application of a semantic segmentation model facilitates precise farmland area estimation and is employed to control UAV flight altitude, ensuring optimal data acquisition. Furthermore, this study optimizes the agricultural subsidy application process by integrating automated inspection and data recording, improving monitoring efficiency and accuracy while minimizing human errors. The integration of UAVs with artificial intelligence (AI) technology provides empirical validation for agricultural applications, advancing precision agriculture while reducing inspection risks and mitigating environmental impacts. Ultimately, this research lays the foundation for a sustainable agricultural inspection system, contributing to the long-term goals of smart agriculture and sustainable development.參考文獻 楊明德、許飪群(2023)。結合無人機遙測, 人工智慧與邊緣運算之 水稻倒伏評估。Journal of Advanced Technology & Management,12(1)。 Eskandari, R.、Mahdianpari, M.、Mohammadimanesh, F.、Salehi, B.、Brisco, B.、Homayouni, S.(2020)。Meta-analysis of unmanned aerial vehicle (uav) imagery for agro-environmental monitoring using machine learning and statistical models。Remote Sensing,12(21),3511。 Gavrilović, M.、Jovanović, D.、Božović, P.、Benka, P.、Govedarica, M.(2024)。Vineyard zoning and vine detection using machine learning in unmanned aerial vehicle imagery。Remote Sensing,16(3),584。 Gupta, S. B.(2023)。Artificial intelligence in smart agriculture: Applications and challenges。CURRENT APPLIED SCIENCE AND TECHNOLOGY,,e0254427-e0254427。 He, K.、Zhang, X.、Ren, S.、Sun, J. (2016)。 Deep residual learning for image recognition, Proceedings of the IEEE conference on computer vision and pattern recognition (第 770-778頁)。 Huang, Y.、Thomson, S. J.、Hoffmann, W. C.、Lan, Y.、Fritz, B. K.(2013)。Development and prospect of unmanned aerial vehicle technologies for agricultural production management。International Journal of Agricultural and Biological Engineering,6(3),1-10。 Kamilaris, A.、Prenafeta-Boldú, F. X.(2018)。Deep learning in agriculture: A survey。Computers and Electronics in Agriculture,147,70-90。 Kerkech, M.、Hafiane, A.、Canals, R.(2018)。Deep leaning approach with colorimetric spaces and vegetation indices for vine diseases detection in uav images。Computers and Electronics in Agriculture,155,237-243。 Kuwata, K.、Shibasaki, R. (2015)。 Estimating crop yields with deep learning and remotely sensed data, 2015 IEEE international geoscience and remote sensing symposium (IGARSS) (第 858-861頁)。 IEEE。 Lelong, C. C.、Burger, P.、Jubelin, G.、Roux, B.、Labbé, S.、Baret, F.(2008)。Assessment of unmanned aerial vehicles imagery for quantitative monitoring of wheat crop in small plots。Sensors,8(5),3557-3585。 Liakos, K. G.、Busato, P.、Moshou, D.、Pearson, S.、Bochtis, D.(2018)。Machine learning in agriculture: A review。Sensors,18(8),2674。 Lottes, P.、Khanna, R.、Pfeifer, J.、Siegwart, R.、Stachniss, C. (2017)。 Uav-based crop and weed classification for smart farming, 2017 IEEE international conference on robotics and automation (ICRA) (第 3024-3031頁)。 IEEE。 Swain, K. C.、Thomson, S. J.、Jayasuriya, H. P.(2010)。Adoption of an unmanned helicopter for low-altitude remote sensing to estimate yield and total biomass of a rice crop。Transactions of the ASABE,53(1),21-27。 Torres-Sánchez, J.、Peña, J. M.、de Castro, A. I.、López-Granados, F.(2014)。Multi-temporal mapping of the vegetation fraction in early-season wheat fields using images from uav。Computers and Electronics in Agriculture,103,104-113。 Tsouros, D. C.、Bibi, S.、Sarigiannidis, P. G.(2019)。A review on uav-based applications for precision agriculture。Information,10(11),349。 Van Klompenburg, T.、Kassahun, A.、Catal, C.(2020)。Crop yield prediction using machine learning: A systematic literature review。Computers and Electronics in Agriculture,177,105709。 Wang, P.、Chen, P.、Yuan, Y.、Liu, D.、Huang, Z.、Hou, X.、Cottrell, G. (2018)。 Understanding convolution for semantic segmentation, 2018 IEEE winter conference on applications of computer vision (WACV) (第 1451-1460頁)。 Ieee。 Xiang, H.、Tian, L.(2011)。Development of a low-cost agricultural remote sensing system based on an autonomous unmanned aerial vehicle (uav)。Biosystems engineering,108(2),174-190。 Zhang, C.、Kovacs, J. M.(2012)。The application of small unmanned aerial systems for precision agriculture: A review。Precision agriculture,13,693-712。 Zhao, B.、Feng, J.、Wu, X.、Yan, S.(2017)。A survey on deep learning-based fine-grained object classification and semantic segmentation。International Journal of Automation and Computing,14(2),119-135。 描述 碩士
國立政治大學
資訊科學系碩士在職專班
111971015資料來源 http://thesis.lib.nccu.edu.tw/record/#G0111971015 資料類型 thesis dc.contributor.advisor 劉吉軒<br>彭彥璁 zh_TW dc.contributor.author (Authors) 蔡孟宗 zh_TW dc.contributor.author (Authors) Tsai, Meng-Tsung en_US dc.creator (作者) 蔡孟宗 zh_TW dc.creator (作者) Tsai, Meng-Tsung en_US dc.date (日期) 2025 en_US dc.date.accessioned 3-Mar-2025 14:29:16 (UTC+8) - dc.date.available 3-Mar-2025 14:29:16 (UTC+8) - dc.date.issued (上傳時間) 3-Mar-2025 14:29:16 (UTC+8) - dc.identifier (Other Identifiers) G0111971015 en_US dc.identifier.uri (URI) https://nccur.lib.nccu.edu.tw/handle/140.119/155992 - dc.description (描述) 碩士 zh_TW dc.description (描述) 國立政治大學 zh_TW dc.description (描述) 資訊科學系碩士在職專班 zh_TW dc.description (描述) 111971015 zh_TW dc.description.abstract (摘要) 本研究旨在建立本研究旨在將無人機與AI模型整合,進行農作物種類辨別以及農田面積估算。使用無人機收集欲辨識之農作物照片及高空農田照片為資料來源。以收集之農作物照片,對物件分類模型(Zhao等人,2017)進行微調,此模型用以農作物種類辨識之用。以收集之高空農田影像,對語意分割(Wang等人,2018)模型進行微調,此模型用於控制無人機飛行高度,飛行高度需達到可辨識所指定之完整農田面積。 系統成效評估是基於以下原則,準確性、效率性、安全性。首先,系統需具備高準確度,以確保農作物種類辨識及農田面積估算的精確性。其次,系統需具備高效性,以快速完成巡查工作,減少人工干預。最後,系統需確保巡查人員的安全,降低人員出行的風險。 本研究的主要貢獻在於提升農田作物巡檢效率,使用無人機替代人工巡檢行為,從而降低人力成本與時間以及巡檢之風險。透過微調的物件分類模型,進行農作物辨識,為農田作物種類提供更可靠的數據支持以及記錄相關影像作為日後查驗。同時,使用語意分割模型的應用使農田面積估算,研究使用語意分割模型進行無人機飛行高度控制。此外,本研究優化了農業補助申請流程,透過自動化檢查與數據記錄,提高巡檢工作的效率與準確性,減少人工作業的誤差。無人機與 AI 技術的整合為農業應用提供了實際資料佐證,推動精準農業的發展,同時降低巡查人員的安全風險,並減少農業活動對環境的影響。最終,本研究為建立可持續的農業巡檢系統奠定基礎,助力實現智慧農業與可持續發展的長遠目標。 zh_TW dc.description.abstract (摘要) The purpose of this study is to integrate drones with AI models for crop type identification and farmland area estimation. Drones will be used to collect photos of the crops to be identified and aerial images of the farmland as data sources. The collected crop photos will be used to fine-tune the object classification model, which is intended for crop type identification. The collected aerial farmland images will be used to fine-tune the semantic analysis model, which is used to control the flight altitude of the drone, ensuring the altitude is sufficient to recognize the specified complete farmland area. The effectiveness evaluation of the system is based on the principles of accuracy, efficiency, and safety. First, the system must have high accuracy to ensure precise crop type identification and farmland area estimation. Second, the system must be efficient to complete inspection tasks quickly and reduce manual intervention. Lastly, the system must ensure the safety of inspection personnel, reducing the risks associated with their fieldwork. This study enhances the efficiency of crop inspection in agricultural fields by utilizing unmanned aerial vehicles (UAVs) to replace manual inspections, thereby reducing labor costs, time consumption, and operational risks. Through the fine-tuning of an object classification model, the system enables accurate crop identification, providing reliable data support for crop classification while recording relevant images for future verification. Additionally, the application of a semantic segmentation model facilitates precise farmland area estimation and is employed to control UAV flight altitude, ensuring optimal data acquisition. Furthermore, this study optimizes the agricultural subsidy application process by integrating automated inspection and data recording, improving monitoring efficiency and accuracy while minimizing human errors. The integration of UAVs with artificial intelligence (AI) technology provides empirical validation for agricultural applications, advancing precision agriculture while reducing inspection risks and mitigating environmental impacts. Ultimately, this research lays the foundation for a sustainable agricultural inspection system, contributing to the long-term goals of smart agriculture and sustainable development. en_US dc.description.tableofcontents 第一章 緒論 1 第一節 研究背景與動機 1 第二節 研究目的與問題 2 第三節 論文架構 3 第四節 研究範圍 3 第五節 預期貢獻 5 第二章 文獻探討 7 第一節 無人機於農業中之應用 7 第二節 AI技術於農業中之應用 8 第三節 無人機與AI技術結合之研究進展 8 第三章 技術方法與系統開發 10 第一節 系統架構設計 10 一、資料收集模組 10 二、數據處理模組 11 三、AI模型分析模組 11 四、無人機行為決策模組 11 五、結果展示與紀錄模組 12 第二節 深度學習模型技術應用 12 一、資料集準備 12 二、資料前處理及資料擴增 17 三、模型訓練 19 四、模型評估 20 第三節 語意分割模型技術應用 34 一、資料集準備 34 二、資料前處理及資料擴增 37 三、模型訓練 39 四、模型評估 40 第四節 農田面積估算 41 第四章 實驗設計與結果分析 45 第一節 實驗設計 45 第二節 實驗評估指標計算 48 一、 農作物辨識 48 二、農田區域辨識 48 三、巡檢任務時間 49 第三節 實驗結果與分析 49 一、 農作物辨識實驗結果 49 二、 農田區域辨識實驗結果 52 三、 巡檢任務時間實驗結果 52 四、 巡檢任務紀錄輸出結果 57 第四節 實驗結果討論 68 一、作物辨識與高度設定 68 二、面積計算誤差分析 68 三、無人機與人員巡檢效率比較 69 第五章 結論與未來展望 70 第一節 研究限制 70 一、影像解析度限制 70 二、 語意切割模型的邊界精度不足 70 三、環境影響未納入考量 70 四、巡檢效率的場景限制 70 第二節 主要貢獻 71 一、 提出動態飛行高度調整方法 71 二、驗證作物辨識高度極限 71 三、提供效率與應用的比較分析 71 四、探索農田面積計算的應用限制 71 五、建立一套模組化巡檢系統 71 第三節 未來展望 72 一、 提升影像解析度與數據質量 72 二、 優化語意切割模型與數據擴展 72 三、 導航與路徑規劃的智能化 72 四、 環境適應能力的提升 72 五、 擴展應用場景與商業化推廣 72 參考文獻 73 附錄 75 zh_TW dc.format.extent 69046518 bytes - dc.format.mimetype application/pdf - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0111971015 en_US dc.subject (關鍵詞) 無人機 zh_TW dc.subject (關鍵詞) 人工智慧 zh_TW dc.subject (關鍵詞) 深度學習 zh_TW dc.subject (關鍵詞) 農作物辨識 zh_TW dc.subject (關鍵詞) 物件分類 zh_TW dc.subject (關鍵詞) 語意分割 zh_TW dc.subject (關鍵詞) UAV en_US dc.subject (關鍵詞) artificial intelligence en_US dc.subject (關鍵詞) deep learning en_US dc.subject (關鍵詞) crop identification en_US dc.subject (關鍵詞) object classification en_US dc.subject (關鍵詞) semantic segmentation en_US dc.title (題名) 農作物空拍影像種類辨識與無人機農檢任務應用效能評估 zh_TW dc.title (題名) Assessment of Crop Aerial Image Classification and Performance Evaluation of Drone-Based Agricultural Inspection Tasks en_US dc.type (資料類型) thesis en_US dc.relation.reference (參考文獻) 楊明德、許飪群(2023)。結合無人機遙測, 人工智慧與邊緣運算之 水稻倒伏評估。Journal of Advanced Technology & Management,12(1)。 Eskandari, R.、Mahdianpari, M.、Mohammadimanesh, F.、Salehi, B.、Brisco, B.、Homayouni, S.(2020)。Meta-analysis of unmanned aerial vehicle (uav) imagery for agro-environmental monitoring using machine learning and statistical models。Remote Sensing,12(21),3511。 Gavrilović, M.、Jovanović, D.、Božović, P.、Benka, P.、Govedarica, M.(2024)。Vineyard zoning and vine detection using machine learning in unmanned aerial vehicle imagery。Remote Sensing,16(3),584。 Gupta, S. B.(2023)。Artificial intelligence in smart agriculture: Applications and challenges。CURRENT APPLIED SCIENCE AND TECHNOLOGY,,e0254427-e0254427。 He, K.、Zhang, X.、Ren, S.、Sun, J. (2016)。 Deep residual learning for image recognition, Proceedings of the IEEE conference on computer vision and pattern recognition (第 770-778頁)。 Huang, Y.、Thomson, S. J.、Hoffmann, W. C.、Lan, Y.、Fritz, B. K.(2013)。Development and prospect of unmanned aerial vehicle technologies for agricultural production management。International Journal of Agricultural and Biological Engineering,6(3),1-10。 Kamilaris, A.、Prenafeta-Boldú, F. X.(2018)。Deep learning in agriculture: A survey。Computers and Electronics in Agriculture,147,70-90。 Kerkech, M.、Hafiane, A.、Canals, R.(2018)。Deep leaning approach with colorimetric spaces and vegetation indices for vine diseases detection in uav images。Computers and Electronics in Agriculture,155,237-243。 Kuwata, K.、Shibasaki, R. (2015)。 Estimating crop yields with deep learning and remotely sensed data, 2015 IEEE international geoscience and remote sensing symposium (IGARSS) (第 858-861頁)。 IEEE。 Lelong, C. C.、Burger, P.、Jubelin, G.、Roux, B.、Labbé, S.、Baret, F.(2008)。Assessment of unmanned aerial vehicles imagery for quantitative monitoring of wheat crop in small plots。Sensors,8(5),3557-3585。 Liakos, K. G.、Busato, P.、Moshou, D.、Pearson, S.、Bochtis, D.(2018)。Machine learning in agriculture: A review。Sensors,18(8),2674。 Lottes, P.、Khanna, R.、Pfeifer, J.、Siegwart, R.、Stachniss, C. (2017)。 Uav-based crop and weed classification for smart farming, 2017 IEEE international conference on robotics and automation (ICRA) (第 3024-3031頁)。 IEEE。 Swain, K. C.、Thomson, S. J.、Jayasuriya, H. P.(2010)。Adoption of an unmanned helicopter for low-altitude remote sensing to estimate yield and total biomass of a rice crop。Transactions of the ASABE,53(1),21-27。 Torres-Sánchez, J.、Peña, J. M.、de Castro, A. I.、López-Granados, F.(2014)。Multi-temporal mapping of the vegetation fraction in early-season wheat fields using images from uav。Computers and Electronics in Agriculture,103,104-113。 Tsouros, D. C.、Bibi, S.、Sarigiannidis, P. G.(2019)。A review on uav-based applications for precision agriculture。Information,10(11),349。 Van Klompenburg, T.、Kassahun, A.、Catal, C.(2020)。Crop yield prediction using machine learning: A systematic literature review。Computers and Electronics in Agriculture,177,105709。 Wang, P.、Chen, P.、Yuan, Y.、Liu, D.、Huang, Z.、Hou, X.、Cottrell, G. (2018)。 Understanding convolution for semantic segmentation, 2018 IEEE winter conference on applications of computer vision (WACV) (第 1451-1460頁)。 Ieee。 Xiang, H.、Tian, L.(2011)。Development of a low-cost agricultural remote sensing system based on an autonomous unmanned aerial vehicle (uav)。Biosystems engineering,108(2),174-190。 Zhang, C.、Kovacs, J. M.(2012)。The application of small unmanned aerial systems for precision agriculture: A review。Precision agriculture,13,693-712。 Zhao, B.、Feng, J.、Wu, X.、Yan, S.(2017)。A survey on deep learning-based fine-grained object classification and semantic segmentation。International Journal of Automation and Computing,14(2),119-135。 zh_TW
