Drowning early warning method based on optical vision analysis

Abstract

The invention provides a drowning early warning method based on optical vision analysis. Firstly, the image is denoised, enhanced and preprocessed to remove uneven illumination through median filtering, histogram equalization and homomorphic filtering, and then the immune genetic algorithm is used to simultaneously detect and track the swimmer and obtain its position Finally, the LSTM network is used to extract the behavior characteristics of the swimmer with time series information, and the characteristics are analyzed through the Softmax layer to obtain the drowning probability of the swimmer, and the judgment of whether to drown is given. The method of the invention can realize intelligent visual perception of swimmers in a swimming pool, and by understanding and analyzing their movement behavior, early warning of drowning status can be given, and the position information of the drowning person can be quickly and accurately given to assist lifeguards in their rescue work.

Description

technical field [0001] The invention belongs to the technical fields of computer vision and graphic processing, and in particular relates to a drowning early warning method based on optical vision analysis. Background technique [0002] Along with the development of society and the raising of people's living standard, the sports of this whole body exercise of swimming are more and more subject to people's welcome. However, because it is carried out in water, beginners often cause choking and even drowning accidents because they cannot breathe freely and move freely. In addition, swimmers may experience muscle spasms, collisions, and physical exhaustion in the water, all of which may lead to drowning events. Once a drowning incident occurs, it will be directly related to the life safety of swimmers. According to the report of the World Health Organization: Drowning has become the third leading cause of unintentional injury death all over the world, and the number of drowning...

AI Technical Summary

Problems solved by technology

[0005] The above-mentioned methods of utilizing detection devices such as sonar have many disadvantages: 1) because most monitoring devices such as sonar sense through the air contained in the human body, if the drowning person's body is full of water, the signal obtained will be low. 2) The sonar method needs to generate sonar pulses at a voltage of several hundred volts, and the relevant safety regulations, there must be no application voltage higher than 12 or 24 volts near the swimming pool water, which is obviously within the scope of safety regulations The internal voltage cannot meet the needs of sonar equipment; 3) The processing scenarios are very limited

However, in the early systems based on visual analysis, the camera was located at the bottom of the pool, and whether the swimmer was drowning was judged by judging whether the swimmer was at the bottom of the pool for a certain period of time. Save time. In addition, when there are too many swimmers, mutual occlusion between swimmers will inevitably occur, making it difficult for the camera at the bottom of the pool to obtain sufficient visual information.

Images