【作者】 张蕊；

【导师】 朱瑞祥； 毛恩荣；

【作者基本信息】 西北农林科技大学 ， 农业机械化工程， 2008， 硕士

【摘要】 粮食生产状况,特别是农作物种植面积,是关系国计民生、社会稳定的重要信息。是制定国家和区域社会经济发展规划、农产品进出口计划、指导和宏观调控种植业结构、预测农业资源综合生产能力与人口承载能力的重要数据源。但长期沿用数据统计的方式,效率低、可靠性差;点状抽样方法需要大量的人力、物力和财力投入,且数据质量也难以保障。本文针对大尺度农作物面积调查存在的问题,在大尺度农作物面积车载调查系统VGGVS(Vehicle&GIS&GPS&Video System)硬件设计的基础上,开发一套地面车载统计辅助系统,以提高大尺度作物分类面积地面调查系统的效率,减少主观人为因素的影响。该系统的实现,具有重要使用价值和广阔的应用前景。本研究在VGGVS硬件设计的基础上,开发了车辆定位模块、导航模块、视频图像采集模块。该辅助系统可提供车辆定位和GIS功能,如查询、缩放地图、漫游、最短路径计算等。并可以辅助VGGVS的图像处理与识别,进行野外采集试验的AVI视频保存、实时提取采集的BMP图像等功能。论文的主要研究的方法与内容如下:(1)本研究将GPS与GIS相结合,采用基于投影的地图匹配技术和Dijkstra算法的路径规划技术,实现了适宜本系统的车辆定位导航和最短路径规划的功能,实现了在GIS底图上进行GPS定位点显示和车辆行驶轨迹线的绘制。(2)为了使大尺度作物面积的计算更加准确,对GPS与时间以及图像的匹配进行了研究。将采集图像与GPS定位信息相匹配,使图像中的作物与其空间位置信息相对应,确立了移动过程中数字化的作物图像位置信息的计算方法。(3)研究了BMP格式图像实时提取的方法。通过倾斜摄影模型计算拍摄图像的长度,依据车速以及摄像头焦距、拍摄角度等信息确定图像采集时间间隔,根据实时采集图像时间间隔,提取单帧作物图像,并以BMP格式存储。该BMP格式图像可直接进行图像处理和图像识别,从而实现在野外采集过程中实时识别作物种类,并计算出作物的种植面积和种植成数。(4)为了野外采集工作的后期检验,研究了保存带有定位信息的作物视频的方法。利用GPS与Video通过时间同一性进行匹配,使采集的视频图像带有精确的时间和位置信息,通过视频回放可以进一步检查、验证作物面积调查的准确性和真实性。实验验证表明,本研究开发的车辆定位模块、导航模块、视频图像采集模块,运行可靠、有效,能够快速准确地实现大尺度作物分类面积地面调查,并具有良好的二次开发性能。更多还原

【Abstract】 The food production condition, specially the crops sown area is the important information related with the national economy, the people’s livelihood and the social stability.It is an important data origin of formulating national and the region socio-economic development plan and the agricultural product import and export plans, instructing and macroeconomic regulating and controlling the crop production structure, and forecasting the agricultural resource synthesis productivity and the population bearing capacity. But the traditional data statistics for a long time is low efficiency and reliability.The punctual sampling method needs massive investment of manpower, physical and financial resource, and the data quality can not be guarantee.In view of existence question of large-scale crop acreage statistical method investigation, this research developed a car-carried ground statistics system which aims at making up shortage of currently large-scale crop acreage statistical method in the national scope,it have an important meaning to raise the efficiency of large-scale crop acreage classification statistical method and reduce the subjective factor influence of the people.Based on the VGGVS hardware design,this research developed vehicle-localization module, video and image acquisition module and vehicle-navigation module. The assistive system has powerful functions ,such as GIS features which provided inquiries, zooming and roaming maps and calculating the shortest path, it could assist the imagery processing and the recognition function to save AVI video files and extract BMP image real-time in wild-collected test also. The main research content is as follows:1.Considering the mobility and localization demand of VGGVS,this research used the map matching technique based on projection and path planning technique based on dijkstra algorithm,realized the vehicle location and navigation function and the shortest path planning function which is suitable for this system,and realized the GPS locating point display and the vehicles travel track drawing on the GIS base map through the match of the GIS and the GPS.2.In order to make large-scale crop acreage investigation more accurate,this research has studied the match of the GPS、the time and the image , matched the gathering image and the GPS localization information to make the crop image correspond with its space position information, and established the accurate positional information computation method of the digitization crops image in the motion process.3.Contrapose the real time and highly efficiency capability of the system crops automatic identification and calculates the crops area, has studied the BMP form image extraction method on real-time. Computed photography image length with oblique photograph model and evaluated the interactive relations of the vehicle speed, camera focal distance, shooting angle information and so on synthetically to confirm the gathering interval, then, extracted single frame BMP form crops image real-time every gathering interval time and carried on the imagery processing and the pattern recognition directly.Thus realized the real-time identifying crops type and calculating the crops area or sown percentage during the open country gathering process .4.For the later period examination after open country gathering work, this study developed the preservation method of crops video which have the localization information, using time identity to match GPS and Video to make the gathering video image have precise time and position information. And make a further inspect to confirm the accuracy and the authenticity of crops area investigation through the video playbacking.更多还原