无人机河湖巡检中RCNN识别算法应用研究  

Research on application of RCNN identification algorithmin UAVS river and lake patrol inspection

作　　者：唐桂荣[1] 许健 钱苏平 付港 钱晓军[3] TANG Guirong;XU Jian;QIAN Suping;FU Gang;QIAN Xiaojun(Taizhou Water Resources Bureau,Taizhou 225300,China;Taizhou Jiangyan Water Resources Bureau,Taizhou 225500,China;School of Computer and Electronic Information,Nanjing Normal University,Nanjing 210023,China)

机构地区：[1]泰州市水利局,江苏泰州225300 [2]泰州姜堰区水利局,江苏泰州225500 [3]南京师范大学计算机科学与技术学院,江苏南京210023

出　　处：《江苏水利》2021年第1期34-38,共5页Jiangsu Water Resources

基　　金：江苏省水利科技项目(2019052)。

摘　　要：随着经济发展与人口增长,带来了河湖治理难度增加、水体污染频发等问题,针对这些情况设计出一种基于区域卷积神经网络算法(RCNN)的无人机河湖巡检图像智能处理平台。通过聚类和构建卷积神经网络提取可能含有河湖污染的区域,利用RCNN对目标区域进行分类和识别,以此作为问题源构建河湖管理的闭环,较好的完成巡检目标。With economic development and population growth,problems such as increasing difficulty in river and lake control and frequent water pollution have been brought about.In response to these situations,an intelligent processing platform for UAV river and lake inspection images based on regional convolutional neural network algorithm(RCNN)was designed.Through clustering and constructing a convolutional neural network,areas that might contain river and lake pollution were extracted,and RCNN was used to classify and identify the target area,which was used as the source of the problem to construct a closed loop of river and lakE-management,and to better complete the inspection target.

关 键 词：河湖治理 RCNN 数字图像处理 

分 类 号：TV85[水利工程—水利水电工程]