首页> 外文会议>Conference on enabling technologies for law enforcement and security >Dynamically reconfigurable vision for intelligent sensing
【24h】

Dynamically reconfigurable vision for intelligent sensing

机译:动态可重新配置的智能感应愿景

获取原文

摘要

This paper introduces our Dynamically Reconfigurable Vision (DRV) technology, and discusses the significant benefits afforded by this technology to law enforcement applications employing unattended sensor networks. Despite improvements in both data compression and transmission technologies in recent years, video bandwidth constraints continue to limit performance for vision systems employed in sensor networks. DRV seeks to reduce, as much as possible the amount of irrelevant spectral information that is collected, and thus to make more effective use of available bandwidth and computational resources than is possible with conventional imaging technology. This is acheived through the intelligent, dynamic allocation of spatial and temporal sensing resources in response to the dynamics of the scene itself. Minimization of irrelevant data in the video processing chain reduces local processing requirements and allows communication of more information in real-time over bandwidth-limited channels. This leads to a reduction in unit power consumption, complexity, cost, andsize - minimizing these system parameters is critical to the performance of unattended sensor systems and networks.We have developed a prototype DRV camera system that demonstrates the significant performance advantages of this technology. Our DRV system employs a real-time reconfigurable CMOS image sensor which supports multiple variable-resolution, independently-configurable windows per exposure and operates in a snapshot capture mode to minimize motion artifacts. Data is output through multiple sensor video ports to minimize readout time and to provide a local contrast enhancement capability. Multiple, time-correlated windows allow a single sensor to fulfill multiple requirements concurrently, such as intruder detection and object tracking. This imager is capable of reconfiguring itself within several microseconds upon demand. Frame-by-frame configurable parameters include frame rate, integration time, and parameters defining each of the windows. Also presented is a design for a very-low-power DRV system that is under development, as well as other planned future enhancements.
机译:本文介绍了我们动态可重构的愿景(DRV)技术,并讨论了这种技术提供了采用无人值守传感器网络的执法应用所提供的显着效益。尽管近年来数据压缩和传输技术都有改进,但视频带宽约束继续限制传感器网络中使用的视觉系统的性能。 DRV寻求减少收集的无关光谱信息的量,从而使传统成像技术更有效地利用可用带宽和计算资源。这是通过智能,动态分配的空间和时间传感资源来实现,以响应场景本身的动态。在视频处理链中的无关数据最小化降低了本地处理要求,并允许在带宽限制信道上实时地进行更多信息。这导致单位功耗降低,复杂性,成本,和大小 - 最小化这些系统参数对无人看管的传感器系统和网络的性能至关重要。我们开发了一种原型DRV摄像机系统,该系统展示了这项技术的显着性能优势。我们的DRV系统采用实时可重新配置的CMOS图像传感器,其支持多个可变分辨率,每次曝光,并以快照捕获模式操作以最小化运动伪影。数据通过多个传感器视频端口输出,以最小化读出时间并提供本地对比度增强能力。多个时间相关的窗口允许单个传感器同时满足多个要求,例如入侵者检测和对象跟踪。此成像仪能够在需求时重新配置在几微秒内。逐帧可配置参数包括帧速率,集成时间和定义每个窗口的参数。还提出了一个设计,适用于正在开发的非常低功耗的DRV系统,以及其他计划未来的增强功能。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号