本文選題：森林火災熱點　切入點：多光譜紅外圖像　出處：《電子科技大學》2014年碩士論文　論文類型：學位論文

【摘要】：通過對森林火災過程中火勢發(fā)展規(guī)律、遙感圖像分析、多光譜紅外檢測機制的探討,完成對森林著火初步模型的建立。論文主體以2009年發(fā)生在澳大利亞墨爾本市威爾遜國家森林公園的森林大火為原型,借助由中科院地球環(huán)境監(jiān)測中心從中國環(huán)境減災光學衛(wèi)星A、B上發(fā)回的圖像熱點進行增強分析,精確的增強圖像有利于監(jiān)測系統后期對熱點發(fā)生火災及擴散方向的預判。本文側重研究利用近地光學衛(wèi)星A、B發(fā)回的紅外多光譜遙感圖像進行分析,總結前人對該類型的數字圖像處理方式,引入對森林遙感多光譜圖像和遙感紅外圖像的新式增強算法,具體工作和結論如下:(1)對遙感多光譜圖像的主成分進行K-L變換(也稱霍特林變換),變換后的各波段圖像矩陣只存在自相關項,協方差系數項均為零,將各分量進行圖像增強,然后將增強后的分量重新合成,分量間的不相關導致各波段均能較好地保持自己的光譜特性,故合成后的圖像能夠保留原圖像的多光譜特性又達到增強圖像目的。通過仿真發(fā)現本算法對原圖像處理后的光譜保持系數K=0.972046,比直接增強方式對光譜特性的保持要優(yōu)12.30%,但該處理方式的峰值信噪比、信噪比、信息熵因素均降低。(2)基于平臺的加權直方圖均衡算法對遙感紅外圖像進行分析,得到的增強圖像與原圖和普通直方圖均衡圖像進行對比,可以從直觀看出該圖像細節(jié)更為清晰。從圖像上進行數據分析,可以得到經本文算法對紅外圖像的處理較普通直方圖均衡算法,在圖像對比度、峰值信噪比、信噪比、信息熵因素均有所提高。峰值信噪比等因素的提高說明本算法處理后的圖像更貼近實際火災熱點分布圖。經過測算本算法熱點區(qū)域對比度相對普通直方圖均衡算法增加約為47.22%,。而對比度的提高,有利于準確判斷火災熱點。
[Abstract]:The development law of fire intensity, the analysis of remote sensing image and the mechanism of multi-spectral infrared detection in forest fire are discussed. The main body of the paper is based on the forest fire that occurred in Wilson National Forest Park in Merben, Australia on 2009. With the help of the hot spots of the images sent back from the China Environmental Mitigation Optical Satellite AHB by the Earth Environment Monitoring Center of the Chinese Academy of Sciences, Accurate enhanced images are helpful to predict the direction of fire and diffusion of hot spots in the later stage of the monitoring system. This paper focuses on the analysis of infrared multispectral remote sensing images from the near Earth Optical Satellite (AMAB). This paper summarizes the previous digital image processing methods and introduces a new enhancement algorithm for forest remote sensing multispectral images and remote sensing infrared images. The specific work and conclusions are as follows: (1) K-L transform (also known as Hotlin transform) for the principal components of multispectral remote sensing images. After the transformation, only autocorrelation items exist in each band image matrix, the covariance coefficients are all zero, and each component is enhanced. Then the enhanced components are recombined, and the uncorrelated components result in each band being able to maintain its own spectral characteristics. Therefore, the synthesized image can preserve the multi-spectral characteristics of the original image and achieve the purpose of enhancing the image. Through simulation, it is found that the spectral holding coefficient K _ (0.972046) of the original image processed by this algorithm is better than that of the direct enhancement method. Excellent 12.30, but the peak signal-to-noise ratio of the processing, The signal-to-noise ratio (SNR) and information entropy are all reduced. (2) based on the weighted histogram equalization algorithm based on the platform, the enhanced image is compared with the original image and the ordinary histogram equalization image, and the enhanced image is compared with the original image and the ordinary histogram equalization image. By analyzing the data on the image, we can get that the infrared image is processed by this algorithm compared with the ordinary histogram equalization algorithm, and the image contrast, the peak signal to noise ratio, the signal-to-noise ratio, and the signal to noise ratio can be obtained. The increase of peak signal to noise ratio (PSNR) shows that the image processed by this algorithm is closer to the actual fire hot spot distribution map. The contrast of the hot spot area of this algorithm is compared with the ordinary histogram equalization calculation. The increase of the method is about 47.221.The contrast is improved. It is propitious to judge the hot spot of fire accurately.
【學位授予單位】：電子科技大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：TP751

【參考文獻】

相關期刊論文 前1條

1 胡海清;魏書精;孫龍;;大興安嶺2001—2010年森林火災碳排放的計量估算[J];生態(tài)學報;2012年17期

相關碩士學位論文 前1條

1 劉麗紅;多光譜圖像融合及其評價方法研究[D];電子科技大學;2012年

，

本文編號：1563480