Font Size: a A A

The Design And Implementation Of Intelligent Video Terminal For The Aged Home Care

Posted on:2018-02-07Degree:MasterType:Thesis
Country:ChinaCandidate:Z Y MaFull Text:PDF
GTID:2348330512488366Subject:Computer technology
Abstract/Summary:PDF Full Text Request
The safety of the aged home care has been widely concerned by the public,as a result of strangers into the house or fall and can not be found in time,resulting in irreversible damage occurs.In order to avoid the repetition of the above situation,this paper designs and development an intelligent video terminal for the aged home care.When the infrared sensor detects unknown objects,the intelligent video terminal is started to carry out the detection of the intrusion detection and the fall detection.If found abnormal,timely voice broadcast to remind,while sending e-mail alerts to users,achieve real-time video transmission function to ensure the safety of the aged at home of personal property.Pedestrian intrusion detection is divided into face detection and face recognition.In the process of face detection,based on cascade classifier of Adaboost featured in Haar,a face detection method is designed.Firstly,the image is preprocessed and then its anti-jamming capability will be enhanced.Secondly,the Haar characteristic is extracted to obtain a cascade classifier with better detection efficiency by Adaboost iterative training.Experiments show that the algorithm can improve the performance of face detection.In the stage of face recognition,the LBP ? SURF and SVM algorithm are used for face recognition.The LBP and SURF operator are applied to extract the features of the image,and then the histogram vector extracted to realize Two-Category classification through the SVM classifier training.Experiments demonstrate that the algorithm has high accurate ratio and lacks sensitivity to illumination,which is fully satisfactory to the real-life scenario to solve the problem of face recognition.Falling detection module proposed is based on Freeman Chain Code falling detection algorithm.GMM(Gaussian Mixture Model)is adopted to detect foreground object.The moving direction and distance of the target contour are recorded by Freeman Chain Code,and then the Euclidean distance method is used to measure the similarity between the current chain code and the falling chain code to determinewhether falling will happen.The time must be limited.The experiments prove that the algorithm proposed in this paper has high detection speed and solves the shortcomings of low efficiency and low accuracy in practical use.Video real-time transmission module is achieving NAT penetration technology by using P2P-based UDP communication method.The client terminal is tethered to the front end through Socket to timely remind users the danger detected by means of the voice and mail,and meanwhile,users can view in real-time VMIs(video monitoring images).Finally,according to the practical requirements of the terminal,the intelligent video terminal based on front end-client is designed and implemented.Experiments prove that this research program is completely satisfying the use requirement of intelligent video terminal.
Keywords/Search Tags:Haar characteristic, LBP characteristic, BURF characteristic, Freeman Chain Code, P2P-UDP
PDF Full Text Request
Related items