Tumble behavior identification method based on posture comprehensive characteristics

Abstract

The invention discloses a tumble behavior identification method based on posture comprehensive characteristics. Comprising the following steps: video data acquisition, human body skeleton diagram extraction, skeleton image data preprocessing, binary image enclosing rectangle construction, LSTM network unit processing, human body effective area calculation, aspect ratio calculation, distance calculation from the human body mass center to the ground, height change rate calculation and tumble occurrence. On the basis of accurately extracting human body skeleton information, the tumble behavior is comprehensively judged by effectively combining various posture characteristics of a human body, so that the detection accuracy is improved; the lightweight Openpose is used, so that the data volume of the lightweight Openpose is relatively light; secondly, an enhanced gradient descent method and an SGD common gradient descent method are used, and the accuracy can be guaranteed while the convergence speed of the model is increased; the trained optimal model can be conveniently deployed to each edge end; the invention is applied to places such as homes and public places.

Description

technical field [0001] The invention relates to the technical field of artificial intelligence, in particular to a fall behavior recognition method based on body comprehensive features. Background technique [0002] The current social reality of rapidly aging population makes the safety and health of the elderly more obvious. The factors that threaten the health of the elderly are not only the internal incentives such as disease and psychology, but more importantly, the fall injury caused by the combination of internal and external factors. The elderly are the main groups that fall and cause death. Studies have shown that 20% of the elderly in my country are seriously injured after falling. Even if the elderly are usually in good health, 17.7% of them still have a serious injury after falling. Especially in the application scenario of the elderly living alone, how to accurately detect the fall behavior in real time and issue an alarm has become one of the urgent research top...

AI Technical Summary

Problems solved by technology

The influence of internal and external factors will lead to a decline in detection accuracy, so it is particularly important to use a method to achieve fall detection

Commonly used methods are generally based on wearable devices, environmental sensors, computer vision, etc. Wearable-based detection methods usually bring serious discomfort to users, especially in the face of elderly people, heavy wearable devices often cause mobility inconvenience; Detection methods based on environmental sensors often require expensive detection equipment to be installed on the floor or wall, with poor flexibility, poor portability, and high false detection rate; with the rapid development of deep learning, people naturally focus on computer vision methods. The method of computer vision can well avoid the shortcomings of the above two methods. Not only that, the computer vision method also has the advantages of low cost, rapid improvement and maintenance, strong portability, and good flexibility. Therefore, we propose A fall behavior recognition method based on body comprehensive features

Images