Human posture recognition method based on skeleton feature point extraction

Abstract

The invention discloses a human posture recognition method based on skeleton feature point extraction mainly to solve the problem on how to recognize human postures using a single camera. The method is implemented by the steps as follows: (1) inputting a video; (2) preprocessing the video; (3) extracting a coarse skeleton; (4) extracting skeleton feature points; (5) classifying the skeleton feature points; (6) performing gesture recognition; and (7) outputting a gesture recognition result. According to the invention, a coarse skeleton is acquired based on the shortest distance from the pixels of a profile graph to the edge, then, skeleton feature points are calculated and classified based on the distance from a chest node, and finally, a gesture is judged according to the relative position information of the skeleton feature points and the chest node. Human actions can be judged without a lot of calculation under the condition that there is only video steam input of a single camera. The human posture recognition method has the advantages of less calculation, low hardware requirement, high accuracy, and strong adaptability.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to a human body posture recognition method based on bone feature point extraction in the field of computer vision. The invention can be used in an intelligent monitoring system to identify abnormal behaviors of human bodies in the environment, and can also be used in conventional cameras to identify human body actions to realize remote control and human-computer interaction. Background technique [0002] At present, human action recognition is mainly based on complex texture calculations or requires the support of other devices besides a single camera. If the action is recognized through texture calculation, most of the color images or grayscale images need to be sampled, learned, and matched. In the application, there is a very obvious problem: the amount of calculation is large, and most of them require long-term calculations due to performance limitations. Obtaini...

AI Technical Summary

Problems solved by technology

If the action is recognized through texture calculation, most of the color images or grayscale images need to be sampled, learned, and matched. In the application, there is a very obvious problem: the amount of calculation is large, and most of them require long-term calculations due to performance limitations. Obtaining results that lead to inability to complete timely remote control

If it is supported by other devices other than a single camera, there are several problems in the application: 1) The cost increases. Whether it is adding a camera or adding an infrared camera and a wearable wireless remote sensor, the cost of the device needs to be greatly increased; 2) The usage scenarios are limited, infrared cameras are not suitable for multi-person situations such as streets, and wearable wireless remote sensors are not suitable for outdoor scenes

The disadvantage of this method is that the number of motion recognition is limited by using the preset motion map, and the difference in shooting angle or contour map extraction will lead to low recognition rate.

The disadvantages of this method are: the amount of calculation for judging actions through feature extraction is large, and the assistance of motion sensors is required to increase the application cost and limit the application scenarios.

The shortcomings of this method are: poor anti-noise data ability; in the case of unsatisfactory extraction of human body contours, it will lead to a high error rate in the connection of bone feature points, so that the correct human body posture cannot be judged

Images