用基于内容的Web图像搜索引擎设计消防电子化道路水源手册

介绍了基于内容的搜索引擎的体系结构和数据库结构设计,说明了检索匹配子系统的工作原理.将语义和图像间建立的初次联系用一张普通的Excel表表示.在消火栓编号、精确地址、辖区中队、能否出水、检查时间、检查人、形状、流量、压力及照片之间建立初次联系,存储在语义库中作为最原始的语义库,用网络蜘蛛自动遍历网站和网页,把所有相关图象或链接抓回来存入本地数据库.独创了先在本管区服务器中进行、后在大管区服务器中抓取的顺序,这种方法使检索精度得到了很大提高,而检索时间相应减少.经过综合测试,在消防电子化道路水源手册中,检索准确率为99.8%,大管区中寻找一幅含有消火栓信息的图片平均时间为2.38 s;局部管区中(对某一地级市)检索一幅图片所用时间为0.8 s.手册设计中,按照火场消防指挥员和战斗员的思维层次采用仿真思维法,语义描述同步考虑三维以上的语言向量,即某一目标靠何有名建筑物、靠何标志物、最近门牌号码等,如某一向量能准确锁定就不用考虑其他向量,如单一向量不能实现锁定就同步交叉锁定,取三维向量的交叉点锁定,成功率为100%.

New method for fire fighting electronic roadway and waterhead handbook design by web-site image search engine

The present paper introduces a search engine system structure and the database concerned on the design of the engine structure, which accounts for the working principle matching with the search - subsystem. The primary semantic and image relation in our database is shown by means of a simple Excel table while the links among the hydrant mark, the accurately address, the popedom lochus, the water skill, time- and name-checking device, as well as the shape, flux, pressure and hydrant image can be searched in a semantic database as the original one. A spider network automatism will be able to facilitate the users to search the whole web station and the web page. We have also created a new procedural method of doing the searching activities. First in the current server and then searching things in the greater server so as to find the comprehensive correlated images or links in the current server database. Thus, it can be seen that much greater precision can be gained by using the present method. At the same time, it is possible to save time for the searching process. Integrating all the testing procedures, the method can help to increase the searching precision to 99.8%. A typical example is that it takes only 2.38 seconds to search out a hydrant photograph in a very large area while 0.8 second is to be taken to get a hydrant photograph in the local district. Furthermore, the paper works out the Fire Fighting Electronic Roadway and Waterhead Handbook for WEB Image Searching Engine with the above-presented method.