基于灰度共生矩阵的震后倒塌房屋遥感信息提取——以2014年云南鲁甸6.5级地震为例

Remote Sensing Information Extraction of Post-earthquake Collapsed Buildings Based on Gray Level Co-occurrence Matrix——Taking Ludian M 6.5 Earthquake in Yunnan Province in 2014 as an example

以2014年8月3日云南省鲁甸6.5级地震龙头山镇房屋破坏为例,采用两种方法对地震前后卫星影像灰度共生矩阵特征参量进行分析研究。结果表明:①震前无房区对比度值小,有房区值大;无房区逆差距值大,有房区值小;无房区熵值小,有房区值大。②震后无房区对比度值较小,基本完好区值较大,倒塌区居中;无房区逆差距值较大,基本完好区值较小,倒塌区居中;无房区熵值较小,基本完好区值较大,倒塌区居中。③地震前后遥感图像灰度共生矩阵特征参量联合分析方法比地震后倒塌区单时相遥感图像灰度共生矩阵特征参量分析方法的提取结果更准确。

Taking the house damage of Longtoushan Town in Ludian after the M 6.5 earthquake on August 3rd,2014 as an example,the characteristic parameters of gray level co-occurrence matrix of satellite images before and after the earthquake are analyzed and studied by using two methods.The results show that:①before the earthquake,the contrast value of the non-housing area is small,and the value of the housing area is large;the inverse gap value of the non-housing area is large,and the value of the housing area is small;the entropy value of the non-housing area is small,and the value of the housing area is large.②After the earthquake,the contrast value of the non-housing area is smaller,the value of the basically intact area is larger,and the collapse area is in the middle;the inverse gap value of the non-housing area is larger,the value of the basically intact area is smaller,and the collapse area is in the middle;the entropy value of the non-housing area is smaller,the value of the basically intact area is larger,and the collapse area is in the middle.③The combined analysis method of gray level co-occurrence matrix feature parameters of remote sensing images before and after earthquakes is more accurate than that of the single-temporal remote sensing images in collapse area after the earthquake.