【作者】 张健；

【导师】 刘毅； 韩宁；

【作者基本信息】 北京林业大学 ， 机械设计及理论， 2010， 博士

【摘要】 视频监测技术是森林火灾预警科研领域中的重要分支。目前我国林火视频监测系统正在从模拟技术向数字技术过渡,但智能化水平不高,普遍存在的若干问题有:1.瞭望塔的选址未经科学计算,监测覆盖率靠经验估算;2.摄像机扫描到火点后不能自动定位,没有成熟的核心算法;3.林区地形环境复杂,导致视频通信误码率高;4.林火视频自动识别研究刚刚起步,误报率较高等。本文为解决嘹望塔选址、火点自动定位、视频信号通讯质量不高这三个问题,在技术与方法上进行了探索,研究主要内容如下:1.通过制作5m精度的数字高程模型,在ArcGIS平台应用空间分析工具和“贪婪算法”定量地计算出最佳监测点数目、位置、覆盖率等关键参数,在有限监测点条件下既符合通视性要求又实现了最大覆盖。研究结果与鹫峰国家森林公园现有的林火视频监测瞭望塔的数量、位置吻合。本项成果是引入通信领域选址方法专门针对林业特殊用途基站的优化选址研究,为林火视频监测点选址提供了理论依据。2.已有文献和工程实验中的多点定位方法虽然精度高,但存在必须交叉覆盖导致增加投资的问题,所以本文提出了“单点定位”的设想,随着数字云台的出现和GIS平台的成功开发使突破这一难题成为可能,作者创新应用高程数据弥补了缺少必要计算条件的不足,利用地形剖面中的几何关系得到了火点定位的收敛判据。考虑到山区高程多变的特殊性,根据GIS(地理信息系统)提供的坡度变化数据获取相应不同的步长间隔数值,可得到坡度与插值点间函数关系式,使搜索效率得到提高。在林火定位系统开发中,使用了包括交汇法在内的多种定位方法,实现了火点定位方法的综合应用。3.采用国际最新的视频压缩编码标准H.264和达芬奇核心芯片关键技术设计了视频服务器,解决了大规模集成电路之间的接口问题,此成果在林业领域应用还是首次。应用上述研究成果对鹫峰国家森林公园进行了林火视频监测系统设计和实验。验证得出单监测点定位模式下,纬度方向定位误差在150m以内,经度方向定位误差在120m以内;双点定位模式下,经度方向定位误差在12m以内,纬度方向定位误差在15m以内。对大门坨、萝卜地监测点为中心的林场辖区进行典型方位的断面提取试验,表明两点均处于较佳观测位置。更多还原

【Abstract】 Forest fire monitoring is an important portion of operations for fire fighting, This paper puts forward a new video surveillance system which based on wireless transmission technology. The main conclusions and achievements are as follows:This paper uses some trains of thought and methods for choosing construction sites employed by the departments of communications and municipal administration to introduce some useful conclusions from the disciplines of topography, botany and forest fire-prevention science into geographic information spatial analysis. Through the space analysis of the data on the slope direction, gradient, elevation and compartments of the Jiufeng National Forest Park, we have decided on the construction-sites of forest fire video monitoring spots and derived a conclusion that the number of video monitoring spots for the Jiufeng National Forest Park should be no more than four. This paper may provide some useful experiences for future forest fire video monitoring project.This paper is for the purpose of studying the forest fire video monitoring system installed on the watch-towers as the basis of hardware by using a data retrieval technology from the single watch-tower on the geographical information technology platform. As is known, automatic fire spot locating is of great significance to the forest fire control and video monitoring system .Therefore, it has become a great need to solve the problem of fire spots locating under the condition of a single monitoring spot. We have tried to use the methods to locate the spots on the terrain profile of DEM layer with the help of the search step interpolation function under different slope gradients. Then the curve fitting method can be used to study the changing characteristics of mountain terrain. By taking into full considerations the judgment of fire spot coordinate under various conditions, it would become possible to reach different convergence criteria. Next software programming can be made based on the aforementioned fire spot locating algorithm to collect and input the multiple data spots in DEM into the software to conduct the fire spot locating verification test. Both field trials and practical application show that the accuracy of this method can meet the actual demand for its designated duties and is expected to play an important role in upgrading China’s smart forest fire emergency command and management.This paper presents a combined program of double-point positioning method and single-point positioning method, in which such advanced software technologies as component GIS, ArcSDE database engine and NET are employed in a comprehensive way to design and develop the forest fire disaster positioning system for Jiufeng National Forest Park, solving the urgent problem of ignition point on-site positioning currently confronted with the forest fire prevention grassroots management.This paper introduces the quantize design principle for realizing H.264 video codec standard and has designed in hardware description language (Verilog HDL) its hardware circuit using the pipeline and look-up table. The hardware modules of the quantize can process in parallel. Its simulation on QuartusII 6.0 platform indicates that the maximum operating frequency of the quantizer has reached 106.69MHz and can thus meet the real-time video data processing requirement.更多还原