Research On Fall Detection Method For Home Elderly People Based On Visual Sensor

Nowadays lots of elderly live alone,but this puts them in danger.This article studied the fall behavior detection method of elderly based on visual sensor.It continuously monitors the elderly living alone and finds the elderly falls down this serious abnormal behavior so that to protect the elderly people life safety.The main work of this article is as follows:In order to solve the problem of indoor complex light,this article studies the Camshift target tracking algorithm that combines LBP and chrominance features.And the algorithm combined with Kalman filter result in the timeliness and accuracy of target tracking were improved.Based on the characteristics of the elderly people movements slow speed and no sudden change of direction.This article proposed a method of fall detection for elderly based on abnormal tracking of moving targets.Based on the clues of loss of target tracking during normal exercise for the elderly people that to study the implicit relationship between mutations in track of motion and falling down abnormal events for the elderly.An elderly fall detection rule classifier based on target velocity,area of target contour and local motion of target is designed.By we analyzing the different postures of the elderly,this article proposes micro action detection algorithm based on HOG feature.Divide the video image into 3×3 areas,then detecting the foreground of the target and locating the area where the elderly people resides.Extracting the HOG features of the area and using the SVM classifier to train and test the sample set.The algorithm realizes the recognition of the micro action of the elderly people and effectively detects the coma behavior of the elderly people in sitting posture.In the analysis of the significant differences of the minimum circumscribed rectangularaspect ratio,center of gravity,and speed characteristics of the target in the elderly sitting,standing up,standing up,falling,normal movement and falling.The SVM fall detection algorithm based on multi-feature fusion was designed and experiments verify the effectiveness of the algorithm for fall detection.