頁籤選單縮合
| 題 名 | 無人機於農作物定位及計數之應用=Icrop Positioning and Counting Using an Unmanned Aerial Vehicley |
|---|---|
| 作 者 | 楊明德; 曾偉誠; 許鈺群; 曾信鴻; | 書刊名 | 中國土木水利工程學刊 |
| 卷 期 | 31:6 2019.10[民108.10] |
| 頁 次 | 頁569-577 |
| 專 輯 | 「智慧測量與工程大數據」特刊 |
| 分類號 | 440.98 |
| 關鍵詞 | 無人飛行載具; 水稻; 植株定位; 計數; Unmanned aerial vehicle; UAV; Rice; Positioning; Counting; |
| 語 文 | 中文(Chinese) |
| 中文摘要 | 為準確評估農作物產量與生長狀態,植株定位(Positioning)技術是評估農作物數量的重要基礎。本研究以水稻為例,建立一套基於UAV影像之植株定位流程,並應用至另外四種農作物計數。本研究植株定位流程分為兩部分,首先將UAV航拍影像建立DSM(Digital Surface Model),最大濾波器(Max Filter)萃取極值獲得興趣區域(Region of Interest, ROI)以定位植株。第二部分以變異數濾波器(Variance Filter)增強影像的植株特徵,利用分類法獲得ROI。將上述兩ROI交集得植株定位成果並盤點數量。本研究水稻植株定位召回率(Recall)為98.3%,精確率(Precision)達99.8%。此技術於西瓜、火龍果樹、香蕉樹與高麗菜上皆能計數,證明本研究能貢獻於多樣農作物盤點。 |
| 英文摘要 | In order to accurately assess crop yield and growth status to achieve precision agriculture, plant positioning technology is an important basis. In this study, an image localization process based on an Unmanned Aerial Vehicle (UAV) was established, and rice seedlings were taken as an example for research and demonstration. For demonstration, the proposed procedure was also applied to the other four different crops, including dragon fruit, banana, watermelon, and cabbage. The proposed procedure in this study integrates UAV operation with Image-Based Modeling (IBM) to a generate Orthoimage and a Digital Surface Model (DSM). Also, spatial Filter and various classification methods enable the plant positioning on the UAV image. The study result confirmed that the proposed procedure is workable by providing a recall rate of 98.3% and precision of 99.8% for seedling positioning and counting as well as dragon fruit, banana, watermelon, and cabbage. In this study, the positioning recall rate and the precision approached 98.3% and 99.8%, respectively, for rice seedlings. The proposed procedure was also applied to watermelon, cabbage, dragon fruit tree, and banana tree to demonstrate the positioning and counting for various crops and plants that shows this research can contribute to crops inventory. |
本系統中英文摘要資訊取自各篇刊載內容。