基于改进YOLOv5的无人机河湖智能巡检系统

打开文本图片集

中图分类号：TP393 文献标识码：A文章编号：1006-8228（2026）01-81-05

Abstract：Curentlythesupervisionofillgalactivitiesinriverandlakewaters（suchasunauthorizedanglingillgalfiing，and unauthorizedswimming）generallyacsprominentisuessuhaslowficiencigcost，icultyinachevingfullcoveragead providingreal-timeesponses.Toaddressthismanagementchallegethispaperproposesaninteligentispectionsysteforrivers andlakesbasedonanimprovedcomputervisionalgorithm.Thesystemutilizesedgecomputingdevicesmountedondronesto performlocalizedandapidanalysisofhig-definitionimagescapturedinealtimeacuratelyidentifingandlocatinguauthozed angling，unauthorizedswimming，andotheractivities.ByoptimizingtheRRT*pathplanningalgorithm，thesystemsignificantly enhancesispectioneficiencyExperimentaldatashowsthatthesystemcanconductround-the-clock，high-precisionandatomated supervisionoftargetwaters，greatlyimprovingthetimelinessandaccuracyofidentifyingillgalactivitiesinwaterareasand demonstrating high practicality.

Keywords：UAV;Edge Computing;ComputerVision;Path Planning

0引言

河湖水域是保障供水安全与防洪安全的重要屏障，然而河湖岸线周围存在违法垂钓、非法捕捞、违规游泳等违法行为[-2]。（剩余5014字）