基于视觉识别的泳池防溺水救援系统设计

打开文本图片集

中图分类号：TP391 文献标识码：A 文章编号：2096-4706（2025）18-0087-07

Abstract：The safety issues of swimming pools have always atracted attention.Inorder to solve the current rescue dilemmaof swimming pool drowningaccident，this paper proposesa designofswimming poldrowning preventionandrescue system based on visual recognition.The system consists of two parts ∵ drowning detection and hull rescue，which can monitor thesafetystatusofpeopleinthe poolinrealtime.Once theswimmerisdetectedinadangerousstate，therescue boat equipped withlife-savingairbags willimmediatelystart，quicklyreachthedrowningpositionandejectinfatableairbagstohelpthemout ofdanger.Thesystemintegrates intellgentmonitoringanddrowningidentification functions，significantlyimproves thesafety factorof theswimmingpool，andpromotes theintellgenceofanti-drowning technologyandtheautomationofrescue.Based on the YOLOv5algorithm，the modelthatcanidentifythehuman body，headsafetystateand lifebuoyisinnovativelytrained. By judging whether the swimmer wears a lifebuoy，swimming speed and posture，andthe frequencyof head safetystate，the drowningcondition isaccuratelyidentified.Compared with thetraditional manualrescue mode，the system improves therescue eficiencyandreduces therisk of human interventionbyintroducingvisual recognitionandautomaticrescue mechanism，which provides a new solution for swimming pool safety management.

Keywords： drowning detection; swimming pool; visual recognition; YOLOv5; OpenCV

0 引言

随着游泳运动的普及与泳池设施的广泛建设，泳池安全问题日益受到关注。（剩余8223字）