• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Instrumentation Science and Technology vol.1 >A Low-cost System to Detect and Track a Moving Target in Dynamic Backgrounds
【24h】

A Low-cost System to Detect and Track a Moving Target in Dynamic Backgrounds

机译：一种在动态背景下检测和跟踪运动目标的低成本系统

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

In order to track the moving target in an automaticial supervision system, we design a track system based on DSP (Digital Signal Processor). For two successive frames within the same backgrounds have great correlativity, the Phase-Correlation algorithm is chosen as the principle algorithm to realize the registration of backgrounds' excursion. The algorithm is based on the Fourier Transform, which has its special advantages in its mature FFT and hardware realization. Before and after the Phase-Correlation algorithm, a set of processing such as auto-selection of processing window, normalization operation and double difference are performed to improve anti-jamming performance. The system is composed of five parts: the video decoder module, the DSP module, the serial port module, the memory module and the power module. The floating-point TMS320VC33 DSP is the main part, which realizes the target-tracking algorithm and controls other modules. To keep track of the control flow, all processes are realized in the assembly language. We experiment the system in an indoor room with daylight lamp in VC++ and in the DSP system. The results indicate that the calculation of the target's excursion can reach a precision of one pixel and the processing time of one frame is less than 20ms and can meet the item's 120ms per frame request.
机译：为了在自动监控系统中跟踪运动目标，我们设计了一种基于DSP（数字信号处理器）的跟踪系统。由于相同背景下的两个连续帧之间具有很大的相关性，因此选择相位相关算法作为基本算法来实现背景偏移的配准。该算法基于傅立叶变换，它在成熟的FFT和硬件实现方面具有特殊的优势。在相位相关算法之前和之后，执行一组处理，例如处理窗口的自动选择，归一化操作和双重差，以提高抗干扰性能。该系统由五个部分组成：视频解码器模块，DSP模块，串行端口模块，存储模块和电源模块。浮点TMS320VC33 DSP是主要部分，它实现目标跟踪算法并控制其他模块。为了跟踪控制流，所有过程都以汇编语言实现。我们在带有VC ++和DSP系统的日光灯的室内房间中对该系统进行实验。结果表明，目标偏移的计算可以达到一个像素的精度，一帧的处理时间小于20ms，可以满足每帧120ms的要求。

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