臺灣博碩士論文加值系統

根據茶業改良場資料顯示，赤葉枯病為茶園常發生的疾病，需要及早發現及時處理，避免擴散。台灣茶園面積較大，人力巡視茶園耗時費力，且不容易發現茶樹疾病及茶園缺株區域。因此，為降低茶樹疾病蔓延、節省人力及智慧管理茶園，本研究利用影像辨識技術建置「以影像辨識為基礎之智慧茶園管理平台」，主要包含三大模組： (1)赤葉枯病辨識模組－利用智慧型手機進行拍攝正常及赤葉枯病茶樹影像各100張，並使用Lab+HSV+過濾植生區域輪廓偵測技術進行辨識，赤葉枯病茶樹影像辨識正確率為91%，赤葉枯病茶樹縮圖影像辨識正確率為85%，正常茶樹影像辨識正確率為100%，正常茶樹縮圖辨識正確率為99%；(2) 茶園生長缺株區域辨識模組－使用無人空拍機拍攝茶園生長正常及缺株區域影像分別為579張及650張，以進行茶園生長缺株區域之影像辨識，結果顯示茶園生長缺株之原始影像正確辨識率為97%、縮圖影像正確辨識率為92%，茶園生長正常影像正確辨識率為95%、縮圖影像正確辨識率為95%；(3)茶樹生長履歷模組－紀錄茶樹採收前、後影像及施肥等資料，建立影像式茶樹生長履歷，作為茶農分析茶樹生長條件的依據，以提高茶葉的生產量及品質，並提升消費者對茶葉安全及品質之信心。

According to the data of improvement field in the tea industry, the cooperblight is a disease frequently occurring in the tea garden. It is necessary to find out and dispose of this disease in time in order to avoid its spread. Therefore, in order to reduce the spread of tea tree disease and save manpower as well as achieve the wisdom management of tea garden, this study used the image recognition technology to establish the " An Intelligent Tea Garden Management Platform Based on Image Recognition ", mainly including three modules as follows: (1) the recognition module of cooperblight – using the intelligent mobile phone to shoot 100 images for each normal tea tree and one attacked by cooperblight, and adopting the technologies of Lab+HSV+Filtering phytogenetic contour detection technology to identify; the correct rate of image recognition for tea tree with cooperblight was 91%, the correct rate of image reduction recognition for with cooperblight was 85%, the correct rate of image recognition for normal tea trees was 100%, and the correct rate of image reduction recognition for normal tea trees was 99%; (2) recognition module of tea garden growth missing plant area – using the unmanned aerial vehicle to shoot normal trees and sparse area in the tea garden with the images of 579 and 650, respectively, so as to identify the tea garden growth missing plant area; the results showed that the correct recognition rate of the original images for tea garden growth missing plant area was 97%, the correct recognition rate of image reduction recognition was 92%, the correct recognition rate of the normal growth images was 95%, the correct recognition rate of image reduction recognition was 95%; (3) the resume module of tea tree growth – recording the data such as imaged and fertilization before and after harvest so as to establish the tea tree growth resume by imaging and take them as the analysis basis of the tea tree growth conditions by the tea grower, which can improve the production and quality of tea, and enhance the consumer's confidence in the safety and quality of tea.

摘要 I
ABSTRACT II
謝誌 III
目錄 IV
圖目錄 VII
表目錄 X
第1章 緒論 - 1 -
1.1 研究背景與動機 - 1 -
1.2 研究目的 - 4 -
1.3 研究流程 - 6 -
1.4 論文架構 - 7 -
第2章 文獻探討 - 8 -
2.1 茶葉產業現況及常見疾病與生長過程 - 8 -
2.1.1 茶葉產業現況 - 8 -
2.1.2 茶樹常見疾病 - 11 -
2.1.3 茶樹生長過程 - 14 -
2.2 影像辨識技術在農業的應用 - 15 -
2.3 履歷平台 - 18 -
2.4 影像辨識技術 - 20 -
第3章 研究方法 - 28 -
3.1 系統架構 - 28 -
3.2 茶樹異常特徵辨識模組 - 30 -
3.2.1 赤葉枯病辨識模組 - 31 -
3.2.2 茶園生長缺株區域辨識模組 - 36 -
3.3 茶樹生長履歷模組 - 40 -
第4章 系統實作 - 41 -
4.1 實作環境 - 41 -
4.2 系統展示 - 41 -
4.2.1 赤葉枯病辨識模組 - 42 -
4.2.2 茶園生長缺株區域辨識模組 - 44 -
4.2.3 茶樹生長履歷模組 - 45 -
4.3 實際結果分析 - 47 -
4.3.1 赤葉枯病辨識結果分析 - 47 -
4.3.2 茶園生長缺株區域辨識結果分析 - 51 -
第5章 結論與建議 - 53 -
5.1 結論 - 53 -
5.2 未來研究方向 - 53 -
5.3 研究範圍與限制 - 54 -
參考文獻 - 55 -
作者介紹 - 59 -

網站文獻
[1] 台灣行政院農業委員會茶業改良場，茶葉相關資料，取自：http://www.tres.gov.tw/show_index.php
[2] 福建茶葉網，取自：http://www.fjtea.cn/
[3] 行政院農業委員會，茶葉進出口相關資訊，取自：https://www.coa.gov.tw/
[4] 茶樹疾病資訊，取自：http://web.tari.gov.tw/techcd/特作/茶樹/茶樹簡介.htm
[5] 產銷履歷農產品資訊網(2016)。履歷資訊及詳細生產記錄，取自：
https://taft.coa.gov.tw/rsm/Code_cp.aspx?ID=1829809&EnTraceCode=10703011569
[6] 聯合國糧食及農業組織，取自：http://www.fao.org/home/zh/
中文文獻
[7] 何思漢(2016)，使用無人空中載具之空拍影像估計森林資訊─以茶園為例，國立臺灣師範大學資訊工程學系研究所，碩士論文。
[8] 林宜貞(2014)，應用灰預測建立茶園面積預測系統之研究，國立屏東科技大學農園生產系所，碩士論文。
[9] 周裕耀(2017)，生物炭對土壤保水性及茶樹成長影響之研究，環球科技大學，碩士論文。
[10] 許賢瑤(2006)，包種茶的起源與名稱的由來。茶訊 841: 2-3。
[11] 蔡志千(2014)，茶樹赤葉枯病之流行病學及非農藥防治，國立臺灣大學植物醫學研究所，碩士論文。
英文文獻
[12] Alattab, A. A., & Kareem, S. A. (2012). Efficient Method of Visual Feature Extraction for Facial Image Detection and Retrieval. Computational Intelligence, Modelling and Simulation (CIMSiM), 2012 Fourth International Conference, 220–225.
[13] Bai, X., Li, X., Fu, Z., Lv, X., & Zhang, L. (2017). A fuzzy clustering segmentation method based on neighborhood grayscale information for defining cucumber leaf spot disease images. Computers and Electronics in Agriculture, 136, 157–165.
[14] Dey, A. K., Sharma, M., & Meshram, M. R. (2016). Image Processing Based Leaf Rot Disease, Detection of Betel Vine (Piper BetleL.). Procedia Computer Science, 85(Cms), 748–754.
[15] Fini, M. R. R., Kashani, M. A. A., & Rahmati, M. (2011). Eye detection and tracking in image with complex background. Electronics Computer Technology (ICECT), 2011 3rd International Conference, 6(2), 57–61.
[16] Guiming, S., & Jidong, S. (2016). Remote Sensing Image Edge-Detection Based on Improved Canny Operator. Communication Software and Networks (ICCSN), 2016 8th IEEE International Conference, 652–656.
[17] Gupta, V., Sengar, N., Dutta, M. K., Travieso, C. M., & Alonso, J. B. (2017). Automated Segmentation of Powdery Mildew disease from Cherry Leaves using Image Processing. Bioinspired Intelligence (IWOBI), 2017 International Conference and Workshop, 1–4.
[18] Islam, M., Dinh, A., Wahid, K., & Bhowmik, P. (2017). Detection of potato diseases using image segmentation and multiclass support vector machine. Electrical and Computer Engineering (CCECE), 2017 IEEE 30th Canadian Conference, 1–4.
[19] Joshi, D., & Pansare, S. (2015). Combination of multiple image features along with KNN classifier for classification of marathi barakhadi. Computing Communication Control and Automation (ICCUBEA), 607–610.
[20] Kalra, A. (2016). A Hybrid Approach using Sobel and Canny operator for Digital Image Edge Detection. Micro-Electronics and Telecommunication Engineering (ICMETE), 2016 International Conference, 305–310.
[21] Liu, X., Zhang, Q., Luan, R., & Yu, F. (2013). Applications of perceptual hash algorithm in agriculture images. Image and Signal Processing (CISP), 2013 6th International Congress, 2(Cisp), 698–702.
[22] Nam, Y., Hwang, E., & Kim, D. (2008). A similarity-based leaf image retrieval scheme: Joining shape and venation features. Computer Vision and Image Understanding, 110(2), 245–259.
[23] Niu, X., Wang, M., & Chen, X. (2014). Image Segmentation Algorithm for Disease Detection of Wheat Leaves. Advanced Mechatronic Systems (ICAMechS), 2014 International Conference, 270–273.
[24] Prasetyo, E., & Adityo, R. D. (2017). Mango Leaf Image Segmentation on HSV and YCbCr Color Spaces Using Otsu Thresholding. Science and Technology-Computer (ICST), 2017 3rd International Conference, 3–7.
[25] Qin, S., Yang, Y., & Jiang, Y. (2013). Gesture Recognition from Depth Images using Motion and Shape Features. Instrumentation and Measurement, Sensor Network and Automation (IMSNA), 2013 2nd International Symposium, 172–175.
[26] Quan, Q., Lanlan, T., Xiaojun, Q., Kai, J., & Qingchun, F. (2017). Selecting candidate regions of clustered tomato fruits under complex greenhouse scenes using RGB-D data. Control, Automation and Robotics (ICCAR), 2017 3rd International Conference, 389–393.
[27] Singh, V., & Misra, A. K. (2017). Detection of plant leaf diseases using image segmentation and soft computing techniques. Information Processing in Agriculture, 4(1), 41–49.
[28] Siswantoro, J., Prabuwono, A. S., Abdullah, A., & Idrus, B. (2015). Automatic Image Segmentation using Sobel Operator and k -Means Clustering : A Case Study in Volume Measurement System for Food Products. Science in Information Technology (ICSITech), 2015 International Conference, M, 13–18.
[29] Tuba, V., Jovanovic, R., & Tuba, M. (2017). Digital Image Forgery Detection Based on Shadow HSV Inconsistency. Digital Forensic and Security (ISDFS), 2017 5th International Symposium, 1–6.
[30] Wang, J., He, J., Han, Y., Ouyang, C., & Li, D. (2013). An Adaptive Thresholding algorithm of field leaf image. Computers and Electronics in Agriculture, 96, 23–39.
[31] Xia, C., Lee, J.-M., Li, Y., Song, Y.-H., Chung, B.-K., & Chon, T.-S. (2013). Plant leaf detection using modified active shape models. Biosystems Engineering, 116(1), 23–35.
[32] Xia, C., Wang, L., Chung, B. K., & Lee, J. M. (2015). In situ 3D segmentation of individual plant leaves using a RGB-D camera for agricultural automation. Sensors (Switzerland), 15(8), 20463–20479.
[33] Zhang, S., Wu, X., You, Z., & Zhang, L. (2017). Leaf image based cucumber disease recognition using sparse representation classification. Computers and Electronics in Agriculture, 134, 135–141.