Automated and Intelligent System for Monitoring Swimming Pool Safety Based on the Edge AI Technique
In Taiwan, drowning remains one of the top three causes of accidental inju
ry-related deaths among children and adolescents. The swimming industr
y faces practical challenges in terms of economics and safety measures.
At NHRI, our team has partnered with the safety management sector to d
evelop the AI Smart Pool Automated Drowning Prevention and Early Warn
ing System. Traditionally, video equipment has been used as passive post-
accident evidence to meet legal requirements. In contrast, the AI Smart Po
ol Automated Drowning Prevention and Early Warning System actively an
alyzes swimmer activities and pre-drowning characteristics to assess safet
y status and trigger automatic alerts. The system offers the following func
tions and advantages.
Adaptability to various pool sizes: The system demonstrates flexibility by i
ntegrating standard monitoring systems and employing computational m
odels for water surface observation. By considering multiple angles, lighti
ng conditions, and color variations, the system achieves highly adaptable
AI monitoring and early warning capabilities.
Assisting lifeguards in maintaining attention and improving judgment effi
ciency: High-resolution images of the entire pool area are captured, overc
oming limitations associated with human visual observation, such as a lim
ited field of view, attention span, and blind spots. The system enables prol
onged observation and indicates specific areas that require judgment, ser
ving as an efficient and accurate reinforcement system.
Prediction of swimming patterns: The AI algorithm employed in the syste
m utilizes various swimming postures, data quantification, calculation, an
d domain knowledge of the swimming process to establish an intelligent
mechanism for pre-drowning, mid-drowning, and post-drowning stages.
The system achieves a recognition accuracy of over 90% in identifying swi
mming patterns, providing real-time and accurate safety monitoring assis
tance.
法人
A deep learning-powered novel artificial intelligence algorithm and syste m to assist in the identification of pneumoperitoneum on abdominal com puted tomography
One Model Fit All: Revolutionary One-Stop Shop System for Predicting Co ronary Stenosis without Normal Database in Myocardial Perfusion Imagin g
Real-time Intelligent Analysis and Recording System for Football Matches
“Kappa” Compact Organic Wastewater Treatment system for Small and Medium sized businesses
Technology maturity:Experiment stage
Exhibiting purpose:Display of scientific results
Trading preferences:Negotiate by self
Coming soon!