An abnormal state detection method for a person needing special nursing based on human body key points

Abstract

The invention relates to an abnormal state detection method for a person needing special nursing based on human body key points. The method is characterized by comprising the following steps: pre-identifying a scene area, a boundary and a specific target, and intelligently maintaining; carrying out human body key point detection on personnel in the nursing scene; performing posture judgment basedon the detected human body key points; and integrating the characteristics, analyzing time-asking parameters of the nursed person in different postures and different areas in the monitoring scene, andcomprehensively judging whether the person is abnormal or not. The method is based on the field of computer vision, the human body abnormal state is detected through human body key point detection incombination with a deep learning method, and the method has the advantages of being low in equipment cost, timely in discovery condition, high in detection accuracy and high in robustness.

Description

technical field [0001] The invention relates to a method for detecting an abnormal state of a person in need of special care based on key points of a human body. Background technique [0002] With the development of the times, society pays more and more attention to vulnerable groups such as the elderly, the sick, and the disabled. Many vulnerable people have to live alone for various reasons, such as elderly people living alone at home, blind people and patients in wards. With the increase of labor costs, many vulnerable people cannot receive effective care and protection. Once they encounter unexpected situations that cause them to fall, faint, lose consciousness or self-rescue ability, and cannot be discovered by others in time, it is likely to endanger their lives. As of 2018, there are about 240 million elderly people and about 85 million people with disabilities in my country. The care of vulnerable people has become an important issue in today's society. [0003] A...

AI Technical Summary

Problems solved by technology

It is impossible to judge the situation where the face is blocked or the face cannot be recognized

[0005] The current gesture recognition method also has the following shortcomings: 1. The gesture is generally judged based on a single frame, and multiple frames are not jointly judged, which is prone to misjudgment

2. The recognition method is relatively simple, and the background light and shadow changes, the movement of furniture and objects will interfere with the recognition

3. The human body posture, time, position and other parameters are not combined and analyzed, and the current human state cannot be well distinguished

4. There is no targeted training for application scenarios. The training picture set is downloaded from the Internet, which does not match the product camera parameters and actual usage scenarios, and the training classification effect is not good.

5. Machine learning and gesture recognition are carried out by extracting feature points in pictures, which are greatly affected by factors such as human clothing, body shape, and environment. The accuracy of gesture classification is not high, and it is also prone to misidentification of human bodies

Images