閔克祥，劉建龍，顧昊，朱慧

　　(江蘇省秦淮河水利工程管理處，南京 210022)

　　摘 要：針對湖泊人工巡查方面存在的巡查效率低下、管理不全面、分工不到位、問題追蹤更新不及時等問題，提出利用無人機智慧巡檢方式，通過無人機沿湖泊岸線航拍提取大量圖像，利用SIFT算法完成圖像無縫拼接，提取圖像主要目標區域，并利用SIFT特征完成目標圖像與原始圖像的比對，從而識別湖泊岸線變化情況。該方法應用在石臼湖固城湖湖泊巡查中，無人機初次飛行時提取了湖泊網格地理數據及建筑物構筑物的信息作為原始數據，完成2期巡檢，覆蓋范圍和識別精度較人工巡查有很大提高。在實際應用中發現在水生物識別的細節和巡查頻次上還有待提高，目前作為湖泊輔助性巡查上優勢明顯，作為一種獨立的巡查方式及在智能分析決策上值得進一步探討。

　　關鍵詞：河湖管理;無人機巡查;SIFT算法;智慧巡湖

　　Explore application of smart lake patrol system in lake patrol management

　　MING Kexiang, LIU Jianlong, GU Hao, ZHU Hui

　　(Management Division of Jiangsu Qinhuai River Hydraulic Engineering, Nanjing210022)

　　Abstract: In view of the problems existing in the artifical inspection of lakes, such as low inspection efficiency, incomplete management, inadequate division of labor, untimely problem tracking and updating, this paper proposes to use the smart inspection mode of UAV to extract a large number of images through aerial photography of UAV along the lake and shore line, use SIFT algorithm to complete image seamless splicing and extract the main target area of the image and use the SIFT feature to compare the target image with the original image, so as to identify the changes of lake shoreline. This method is applied to the patrol inspection of Shijiu Lake and Gucheng Lake. During the first flight of UAV, the grid geographic data of the two lakes and the information of buildings and structures are extracted as the original data to complete the two phases of patrol inspection. Compared with manual patrol inspection, the coverage and recognition accuracy are greatly improved. In practical application, it is found that the details of water biometric identification and patrol frequency need to be improved. At present, as an auxiliary patrol of lakes, this method has obvious advantages. As an independent patrol method and in intelligent analysis and decision-making, this method is worthy of further discussion.

　　Key words: riverand lake management; UAV patrol,; SIFT algorithm; smart lake patrol