Water rescue system based on space-time big data analysis and drowning alarm prediction method

Abstract

The invention discloses a drowning alarm prediction method based on space-time big data analysis. The method comprises the steps: building a database of different human body heights and different water area scenes, employing a multi-task deep convolution neural network model for training according to the database, and obtaining a deep learning-based water area detection model and a deep learning-based human body height detection model; acquiring image data, extracting a human body outside a water area in the image based on a deep learning water area detection model, calculating the size of theimage occupied by the human body, and obtaining the actual height of the human body based on a deep learning human body height detection model; if the actual height of the human body is lower than the height threshold, marking the human body as a juvenile, otherwise, marking the human body as an adult; and dividing early warning levels according to the height detection result, and performing early warning according to the divided early warning levels. Alarm condition prediction is carried out by combining waterside human body height detection and drowning person detection, a two-stage early warning and alarming linkage scheme is developed, an early warning level is set, and real-time alarming is carried out, so that timely and effective rescue of the drowning person is realized.

Description

technical field [0001] The invention relates to the technical field of water rescue, in particular to a water rescue system based on spatio-temporal big data analysis and a drowning warning prediction method, which applies big data technology and deep learning technology to the water rescue system. Background technique [0002] my country has a vast water area, and there are a large number of artificial lakes, reservoirs, lakes, rivers, etc. There are problems such as inadequate safety measures and dead spots in monitoring blind spots. People who fall into the water or jump over bridges happen from time to time, which will cause drowning accidents, especially in summer vacations. The season of high incidence of minor drowning. When people fall into the water, they usually call for help passively and respond afterwards. Traditional rescue methods are inefficient, and accidents to rescuers caused by launching rescues are not uncommon, causing huge social impact. [0003] In or...

AI Technical Summary

Problems solved by technology

However, the existing water rescue technology mainly focuses on hardware design. At present, there is still a lack of nationwide drowning information and water rescue big data support platforms, and there is no effective and efficient prediction method for drowning alarms.

In order to achieve this goal, it is necessary to identify the human body and age. The existing technology is usually based on face or height detection, and the face will be affected by the observation angle or water, while the technology for identifying human body height is usually carried out by reference surface calibration. measurement, which will be limited in practical applications in the field of water rescue

Images