Intelligent wheelchair static gesture identification method

Abstract

Disclosed is an intelligent wheelchair static gesture identification method. The method includes the steps of 1) collecting scene depth information through Kinect, 2) separating gestures in the scene depth information according to the depth information, 3) extracting the feature vector of the gestures by using normalization central moment, and 4) conducting gesture identification through a DAGSVM classifier according to the feature vector of the gestures extracted in the step 3). According to the intelligent wheelchair static gesture identification method, in the generating process of the DAGSVM classifier, the between-class distance and the class standard deviation of each SVM classifier are calculated, SVM classifiers are arrayed in the descending order according to the between-class distances, and the SVM classifier with the largest between-class distance is chosen to be used as a root node classifier of the DAGSVM classifier. Similarly, the SVM classifiers with the largest between-class distance at the positions of the rest nodes can be chosen. The phenomenon of error accumulation can be effectively reduced, the correctness of a gesture recognition result is improved, and the safety and the reliability of an intelligent wheelchair man-machine interaction system are furthest guaranteed.

Description

technical field [0001] The invention relates to the field of intelligent wheelchairs, in particular to a gesture recognition method for intelligent wheelchairs. Background technique [0002] At present, the aging of the global population age structure is becoming more and more serious. The number of elderly people aged 60 and above in my country has reached 132 million, accounting for 10% of the total population of the country, and it continues to grow at an average rate of 3.32% every year. Among them, about 50% of elderly patients need to provide nursing services. In addition, there are about 60 million people with various types of disabilities in my country, accounting for about 5% of the total population of our country. Therefore, a total of about 100 million people in our country need to provide nursing care. The physical and physical defects of the elderly and the disabled bring a lot of inconvenience to themselves and their families, and also seriously hinder them f...

AI Technical Summary

Problems solved by technology

Although some of the gesture-based human-computer interaction methods mentioned above are natural in form, they are not very convenient for users because they need to be implemented in conjunction with different parts of the human body.

Images