典型震害遥感图像的模型分析

Pattern analysis of remote sensing imagery for some typical earthquake damages

实际震害复杂多样，反映在遥感影像上很难用某种固定的模式进行概括分析，因而震害遥感影像的解释至今还停留在人工判读水平上，难以做到快速、准确．地震应急响应要求尽量在最短时间内了解灾情以救灾抢险，利用遥感技术和图像模式识别技术是快速确定震害乃至灾情的重要途径．本文以１９７６年中国唐山地展震后航空遥感图像为主要样本图像，参照国内外其它地区震害遥感图像，进行了震害的分级，概括了相应的遥感图像特征，在此基础上，建立了提取震害样本图像各种特征参量的数学模型，初步分析了表征各种震害样本图像的参量组合及其阈值范围的确定方式．

The earthquake damage is very complicated, so that to analyze broadly earthquake damage pattern of remote sensing image is difficult. At present, some person could read earthquake damage from remote sensing image by means of 'eyes', but the procedure is very slow and not accurate. Post-earthquake emergency response demands to know disaster information at very short period, while modern remote sensing and image pattern recognition technology are important approachs to obtain the. earthquake damage information. In this paper, a grading system of earthquake damage is built, and the corresponding features of remote sensing image are summarized mainly from the aerial remote sensing image of the great Tangshan earthquake in 1976 as a sample image. Besides, some mathematical patterns which can draw earthquake damage feature of sample image are structured and every parameter group and its threshold interval which can represent different seismic damage grade are studied.