Research Of Key Frame Extraction In Project Supervision System

At present, the law of supervision has become an important part of construction project. In China, the implementation of the lay of supervision is a reformation of traditional project management system. By now, the supervision of construction project mainly depends on the work style that experienced supervisors walk around the construction site and supervise the project. Supervisors with rich experience have become a scarce resource for construction supervision, which limits the size the supervision companies and thereby limits the development of whole supervision industry.Project Supervision System aimed at providing information technology support to the construction project. Supervision companies have to preserve a large number of supervision records. This will help to ensure the successfully completion of the construction project, and also avoided the dispute among the project owner, the construction company and the supervision company. Now, with the development of computer technology and multimedia technology, image and video has become the most effective information carrier. Video data, relative to the traditional form of the text information, has more rich content which can be expressed in the form of a more intuitive way. Therefore, video information is the most appropriate information form for the Project Supervision System.Project Supervision System is the combination of the real-time monitoring, video Conference, and traditional management information system. Through real-time monitoring functions of the system, video data captured by supervisors will be stored in the database. These video data will serve as the record of supervision work. During the system working, there will be a lot of video data. These video data for the project supervision system must be effectively processed. Video-data-processing mainly focuses on how to do retrieval fast and effectively. As a result, the concept of Content-Based Video Retrieval was established. It was developed from the content-based image retrieval. The Content-Based Video Retrieval uses color texture, shape and characteristics of the movement to establish indices and retrieval video data based on these indices. This makes the process going in a way more like human feelings. Video content-based retrieval techniques applied to Project Supervision System is the innovative points of this paper.Key frame is the mostly used video data index. Key frame extraction is the most important step in Content-Based Video Retrieval. We studied the existing key-frame method, analyzed their characteristics based on the nature of Project Supervision System, selected the most suitable method for extracting key frame and finally designed retrieval module structure.