The Multi-source Remote Sensing Images Fusion For Post-earthquake Collapsed Houses Research

The earthquake disaster happens suddenly. It is one of serious natural disasters and distributes very broadly in our country.The earthquake mechanism is complex. We can't forecast accurately at present When occurs in the dense population city or the towns.It's easy to cause the serious personnel casualty and the economic loss. Therefore after the earthquake occurs, we fast gains accurate disaster situation and adopts the effective rescue work which is most effective method to reduce the disaster loses at present. The post-earthquake collapsed houses is suffered injury the personnel economic loss degree which under the degree attribute earthquake function the ground movement intensity size as well as the earthquake disaster create. Face the earthquake collapsed houses information extraction by multi-source remote sensing data fusion, which is a important process of remote sensing data processing in the earthquake disaster monitors. The multi-source remote sensing phantom fast fusion can obtain the high quality, high spatial resolution remote sensing images and enhance the extraction precision of collapsed house information in order to make up the sole remote sensing data gain information ability the insufficiency.According to different remote sensing images and different application purpose select different fusion method although many remote sensing image fusion methods can be used. From this aspect, article analyzed existing general fusion methods, selected and designed the multi-source remote sensing images fusion process and methods of town post-earthquake collapsed houses in order to meet the needs of real-time disaster emergency.Firstly, this article compared the existing fusion methods and studied to improve SAR and the optical image fusion algorithm. From the color transformation, the arithmetic operation and picture inversion to introduce the existing commonly used fusion methods so that we can clear about the good and bad points of each method. At the same time, considered the existing majority general fusion methods are proposed based on between the optical images fusion. This article made the fusion algorithm improvement in SAR and the optical image fusion. This method effect is ideal after experiment. It applicable to the earthquake collapsed houses radar image and multispectral optical image fusion.Then, studied key technologies of the post-earthquake collapsed houses information of multi-source remote sensing images fusion. It includes fast fusion flow design, image preprocess, fusion methods and fusion effect evaluation selection, according remote sensing image data combination to recommend concrete fusion method based on concrete commonly and so on. The multi-source remote sensing data fast fusion of town post-earthquake collapsed houses mainly manifests that the disaster risk regionalization frail areas are first fused. According to application precision requests to select the proper attention to both effect and the speed fusion algorithm. The idea of fusion method selection is that taking the Sichuan earthquake disaster area as study area, take the town collapsed houses as the object of study, using the multi-source remote sensing images for the data source (including optical image and radar image), using existing fusion methods for comparative study and repeated test. According to standard that the faster, enhancement effect is good, the spectral characteristic maintenance effect superior, the spatial detail information is improved. Preliminary selected 6 fusion methods which is the color standardization to transform (Brovey), HSV, high-pass filter(HPF), IHS, the principal components method of transformation (PCA). According to the different data combination, recommends one most superior fusion method from these 6 methods. From the qualitative and quantitative to evaluate fusion effect. The qualitative evaluation includes image spatial resolution, clarity, the collapsed houses detail and the edge readability, linear features prominent degree, the resolution enhancement of effect limit magnification. From the spectrum characteristic, distinguished around fusion images similar features whether has close or the same color, whether highlight the collapse of housing information, local color difference (spectrum characteristic distortion degree), overall color effect difference and so on. The quantitative evaluation index includes the average value, the standard deviation, the entropy, the average gradient and the spectral distortion, root-mean-square error. Mean value reflects image overall brightness. The standard deviation and the entropy main reflect strong and the weak of image detail information. The spectral distortion and the root-mean-square error main reflect image spectrum maintenance effect. According to the concrete commonly used remote sensing image data combination, recommends the concrete fusion method based on the high, medium, low resolution, as well as the radar and the optical image.Finally, carried on multi-source image fusion test of the town post-earthquake collapsed houses information. It includes time response of fusion method, as well as the concrete data fusion effect comparison and evaluation. Further confirmed the recommendation fusion methods are practical.This article established multi-source images fast fusion flow of town Post-earthquake collapsed houses and studied suitable multi-source images methods for town collapsed houses information enhancement which has important significance on earthquake disaster loss assessment, disaster relief and disaster reduction, the decision.