Behavior detection and effect evaluation method and device for rescue training

Abstract

The invention relates to the technical field of image recognition, in particular to a behavior detection and effect evaluation method and device for rescue training. The method comprises the steps of obtaining a to-be-recognized image; inputting the to-be-recognized image into a VGG-19 convolutional neural network to obtain a feature image sequence; inputting the feature image sequence into an Open-pose network model, identifying 15 skeleton points of the to-be-identified image through the Open-pose network model, and converting coordinates of the 15 skeleton points into a digital sequence; normalizing the digital sequence, calculating limb length and angle parameters, and transmitting the limb length and angle parameters into a three-layer full-connection layer to obtain a pre-estimated posture type; and obtaining a pre-stored digital sequence of sample skeleton point coordinates corresponding to the pre-estimated attitude type, and judging whether the to-be-recognized image reaches the standard or not according to the sample digital sequence and the digital sequence corresponding to the to-be-recognized image. The device is low in manufacturing cost, high in efficiency and capable of being used in a large range.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a method and device for behavior detection and effect evaluation oriented to rescue training. Background technique [0002] Emergency rescue refers to activities and plans for prevention, preparation, response and recovery of sudden and destructive emergencies. Emergency rescue plays an important role in the event of a disaster, escorting the safety of people's lives and property. Earthquakes, mine collapses, fires, floods and other disasters have always threatened the safety of people's lives and property. According to the statistics of natural disasters in the first quarter of 2020 released by the Ministry of Emergency Management, various natural disasters including fires, floods, and earthquakes have caused a total of 5.048 million 42 people died and disappeared, 24,000 people were urgently relocated; more than 1,100 houses collapsed, and 133,000 houses were damaged...

AI Technical Summary

Problems solved by technology

However, using hardware equipment for training has many disadvantages. Heavy sensors will affect the actions of the trainees and limit the actions of the trainers. Moreover, the detection indicators of the hardware equipment are single, and a variety of different equipment will be required during the training process. High and has a certain loss

In addition, some studies have pointed out that virtual technology is used to achieve training, such as using virtual reality (VR) technology to simulate real disaster scenes, so that the trainees can be on the scene, but the VR effect with better effects is very expensive, and it is not suitable for widespread use

Images