The video monitoring of outdoor sites is a demanding task that is commonly tackled by having security guards look at arrays of CCTV monitors. Experience shows that this is largely ineffective, both as a detector and a...
详细信息
The video monitoring of outdoor sites is a demanding task that is commonly tackled by having security guards look at arrays of CCTV monitors. Experience shows that this is largely ineffective, both as a detector and a deterrent. However, modern digital imaging systems can solve both these problems by maintaining constant vigilance 24 hours a day. These systems can be versatile and can operate in several different modes, video motion detection, video nonmotion detection and incident capture, thus providing a flexibility of application environment. By basing these systems on powerful PC technology the end user benefits from a large range of facilities at relatively low cost. In particular, it is possible to have low cost frame storage and high performance communications over telephones, ISDN or Ethernet. image sequences both prior to and after an event can be stored and transmitted. Archiving and retrieval of events can be done efficiently through standard databases. However, in order that such systems be operationally viable it is essential that the detection algorithms be smart enough to reduce the number of false alarms to virtually zero. Most of the discussion concerns technology that is currently available and in everyday use: the author uses the ASTRAGUARD product as a specific example of such a system.
The amount of video information used for scientific purposes is increasing as new ocean observing systems are deployed. To date, underwater video has primarily been collected by SCUBA divers, remotely operated vehicle...
详细信息
The amount of video information used for scientific purposes is increasing as new ocean observing systems are deployed. To date, underwater video has primarily been collected by SCUBA divers, remotely operated vehicles (ROV) and human-occupied submersibles. However, large quantities of video information are likely to be collected by other platforms in the future, such as cabled observatory systems (for example, the VENUS, MARS and NEPTUNE projects) and autonomous underwater vehicles (AUV). This video information is extremely useful for science but locating events or objects of interest from hundreds or thousands of hours of video requires a system for effectively managing video annotation information. The Monterey Bay Aquarium Research Institute (MBARI) uses high-resolution video equipment to record over 300 remotely operated vehicle (ROV) dives per year. Over the past sixteen years, 16,000 videotapes (equates to over 13,500 hours) have been archived and managed as a centralized institutional resource. MBARI has developed a software and hardware system, video Annotation and Reference System (VARS), to facilitate the creation, storage, and retrieval of video annotations based on ROV dive tapes. The VARS components reference a knowledge database of over 3500 biological, geological and technical terms. This hierarchical information allows for consistent and rapid classification, description, and complex querying of objects observed on video. MBARI currently has over 1.8 million annotations stored in the VARS system. These annotations have been used in over 200 professional publications and presentations, in addition to numerous education and outreach projects. These publications span multiple sub-disciplines within biology, geology and chemistry. An operational VARS suite consists of a relational database, a videotape recorder (VTR) that supports the Sony 9-pin protocol over RS-422, and the three VARS applications: Knowledgebase, Annotation, and Query. The relational d
databases are increasingly being used to store multi-media objects such as maps, images, audio and video. storage and retrieval of these objects is accomplished using multi-dimensional index structures such as R*-tree...
ISBN:
(纸本)9780897919951
databases are increasingly being used to store multi-media objects such as maps, images, audio and video. storage and retrieval of these objects is accomplished using multi-dimensional index structures such as R*-trees and SS-trees. As dimensionality increases, query performance in these index structures degrades. This phenomenon, generally referred to as the dimensionality curse, can be circumvented by reducing the dimensionality of the data. Such a reduction is however accompanied by a loss of precision of query results. Current techniques such as QBIC use SVD transform-based dimensionality reduction to ensure high query precision. The drawback of this approach is that SVD is expensive to compute, and therefore not readily applicable to dynamic databases. In this paper, we propose novel techniques for performing SVD-based dimensionality reduction in dynamic databases. When the data distribution changes considerably so as to degrade query precision, we recompute the SVD transform and incorporate it in the existing index structure. For recomputing the SVD-transform, we propose a novel technique that uses aggregate data from the existing index rather than the entire data. This technique reduces the SVD-computation time without compromising query precision. We then explore efficient ways to incorporate the recomputed SVD-transform in the existing index structure without degrading subsequent query response times. These techniques reduce the computation time by a factor of 20 in experiments on color and texture image vectors. The error due to approximate computation of SVD is less than 10%.
Addressing the growing challenges posed by increasing vehicles and stringent traffic violations, many vehicle number plates go undetected while passing through toll booths and traffic lights. Thus it creates the need ...
详细信息
ISBN:
(数字)9798350355338
ISBN:
(纸本)9798350355345
Addressing the growing challenges posed by increasing vehicles and stringent traffic violations, many vehicle number plates go undetected while passing through toll booths and traffic lights. Thus it creates the need for a vehicle number plate recognition system. Existing systems employing ANPR algorithms utilize MATLAB and K-Nearest neighbour algorithms for image processing. Drawbacks include limitations in adapting to dynamic environments, Real time recognition, handling multiple vehicles, slow processing speeds, incapability of multiple object detection in single frame. The paper focuses on Automatic Number Plate Recognition (ANPR), employing specialized cameras and advanced image processing and using YOLOv8 for object detection and extracting license plate information from video feeds using EasyOCR. The noisy data generated during character recognition is filtered out using Regex based algorithm. The system integrates with centralized databases for real-time vehicle information access for users. The same database is used to automate e-challan updation, contributing to traffic management by swiftly issuing fines for rule violations. The proposed video-based processing system achieves number plate detection using YOLOv8 and storage of vehicle number plates in a centralized MongoDB database with an accuracy achieved of 96% in license plate detection and reduces noisy data using algorithm. It also enables real-time data retrieval through a user interface using Reactjs as frontend and is also connected to the MongoDB database for precise and user-friendly access in modern traffic systems.
暂无评论