Font Size: a A A

Design And Implementation Of The Armed Police Camp Monitoring System Based On Panoramic Image

Posted on:2017-02-26Degree:MasterType:Thesis
Country:ChinaCandidate:J ChenFull Text:PDF
GTID:2428330569498688Subject:Electronic and communication engineering
Abstract/Summary:PDF Full Text Request
Camp monitoring is an important research direction in image analysis and understanding field.Based on the current situation of Chinese armed police forces,this thesis aims to propose a method to upgrade current equipment with an affordable price and realize more functions like panoramic image,moving object detection,moving object tracking and intelligent alarm.The main work is as follows:1.It was gave out the analysis and contrast the pros and cons of mosaic algorithms.Then based on the practical use,it proposed a suitable system blueprint for best matching and fusion algorithms.Secondly,the original phase stitching algorithm is further simplified with local characteristics,and improve the system operation speed.Thirdly,it proposed an overall system framework with platform control method.2.An improved moving object detection method based on three frame difference is proposed,which can effectively improve the integrity of moving object detection.Firstly,For the deficiency of the frame difference method,the fast segmentation method based on graph theory is introduced.Secondly,by defining the fusion rule,the regions of the frame difference and divided regions from segmentation are complemented by each other,and the complete contour of the moving object is detected.Third,to further verify the effectiveness of the proposed algorithm,we select the open video test datasets as experiment data,then test the proposed method with frame difference algorithm and improved algorithm.The experimental results prove the effectiveness of the proposed algorithm.3.An improved object tracking algorithm based on meanshift is proposed.Due to the shortcomings of the traditional camshift algorithm,we combined camshift algorithm with Brisk feature points,firstly,Brisk feature points is extracted from the object region of every frame.Secondly,in the camshift algorithm,when the search box size or moving distance is bigger than a pre-defined threshold,resize the box and relocate it by matching Brisk points.4.The video surveillance system based on the panorama mosaic is designed and realized.The system requirement is analyzed,and the framework of system is designed including hardware and software aspects.in the end,the system working process are introduced in detail.
Keywords/Search Tags:video surveillance, image mosaic, moving object detection, moving object tracking
PDF Full Text Request
Related items