基于LSSVM的火灾预警系统研究.docxVIP

作者：大月亮 MACROBUTTON MTEditEquationSection2 SEQ MTEqn \r \h SEQ MTSec \r 1 \h SEQ MTChap \r 2 \h 基于LSSVM的火灾预警系统研究 摘 要：随着科学技术的快速发展，城市化已经越来越快，城市化的发展带来了很多便利的同时也存在着人口密集居住，生产生活集中，建筑封闭且集中灯特点。以此导致火灾发生的频率越来越高。然而火灾存在着许多不确定因素，通过传统的线性模型很难对火灾这类非线性问题进行准确预测。 本文通过分析和对比各种算法之间的优劣，决定采用LSSVM的预测算法，同时我们的图像处理是通过RGB-HIS联合对图像进行特征提取，通过LSSVM对特征提取出的数据进行训练，再通过对预测算法的参数优化，一种新型的用于非线性火灾的预测模型就被搭建出来了，通过最后的实验结果和仿真测试，LSSVM预测算法能够非常准确的对火灾进行预测，于此同时预测的精准度大幅提高。 关键词：火灾预测 LSSVM 线性 非线性  Research on fire warning system based on LSSVM Abstract：With the rapid development of science and technology, urbanization has been more and more rapid, the development of urbanization has brought a lot of convenience, but also there are densely populated living, concentrated production and life, closed buildings and centralized lights. This has led to an increase in the frequency of fires. However, there are many uncertain factors in fire, so it is difficult to predict accurately the nonlinear problems such as fire by the traditional linear model. In this article, through analyzing and comparing the advantages and disadvantages between various algorithm, has decided to adopt the LSSVM prediction algorithm, at the same time our image processing is through the RGB - HIS coalition for image feature extraction, feature extracted data through LSSVM training, again through the prediction algorithm of parameter optimization, a new kind of nonlinear fire prediction model was built for, by the final experiment results and simulation tests, LSSVM prediction algorithm is very accurate forecast fire, at the same time prediction precision greatly improve. Key words: fire prediction LSSVM Linear nonlinear 目 录 TOC \o 1-3 \h \z \u 1 绪论 1 1.1 研究背景及意义 1 1.2 国内外研究现状与发展趋势 1 1.2.1 国外研究现状 1 1.2.2 国内研究现状 2 1.3 课题的研究内容及要求 2 2 算法基本概念 3 2.1 最小二乘法 3 2.1.1 最小二乘法的由来 3 2.1.2 线性最小二乘的原理 3 2.2 支持向量机 4 2.2.1 SVM的由来 5 2.2.2 SVM算法 5 3 基于LSSVM火灾预警 7 3.1 火灾预测原理 7 3.2 火灾预测的LSSVM算法 7 3.3 火焰判别 9 3.3