Research And Implementation Of Earthquake Damage Assessment Methodology Based On Remote Sensing Images

Earthquake disaster is one of the most severe natural calamities in the world, especially in China. Therefore, it is of importance to organize and implement emergency activities after the disaster happened. Earthquake damage loss assessment is the key task in emergency, which is to estimate the damage loss by surveying and analyzing the damage information. At present, the assessment is mainly carried out by disaster locale estimation, but it needs several days to obtain assessment result after the estimate makers arrived at the disaster area. With the appearance of high resolution sensors, it is becoming a trend to use remote sensing (RS) technique to survey the earthquake damage. This situation suggests that it is conceivable to do a quick damage loss assessment using RS images to support the succor decision-making before the locale work finishes.The assessment based on RS images serves emergency and estimates the macroscopical damages, the basis of which is assessment models. The estimate contains two modes, one is based on damage index, and the other is based on image classification. The former gets the assessment result in 2 hours after disaster images are enhanced. The latter is computer assistant and much more accuracy than the first mode, but it needs longer time about 6 hours. The damage index is given by visual interpretation, and then is converted into equivalent ground damage index by which the ground intensity distribution could be produced. Because intensity is relative to construction vulnerability, the computation parameters, buildings'damage rate and loss rate, will be acquired from the experiential vulnerable matrix of different region. Image classification to extract the damage information is based on computer digital image processing technology. When the extraction work is finished, sorted and graded damage data are gained and used to calculate the damage rate. As the two different methods, there are also two loss assessment procedures which are consisted of by estimate models. Models'parameters come from the basic data or semi-finished result. The data not only contains the remote sensing images, but also includes the social and economic information. It is difficult to collect the lifelines'damage information from remote sensing images right now, so an indirect method is proposed to assess the earthquake direct damage loss. After the analysis of 115 earthquake examples, a linear relation is established between construction earthquake direct damage loss and earthquake direct loss. The relation is used to do computation of seismic direct losses. While the direct loss is obtained, one modified model is used to correct the result. The modified model has multiple factors that include the magnitude, focus depth and economic development level. All values of those factors could be count by using the non-line regression method.To exert the remote sensing ability in earthquake emergency, it is very necessary to develop a professional system. The entire assessment course is divided into 3 steps which are preparation, damage information extraction and loss assessment. In the first step, the main task is to collect the basic data. The second step is to use different mode to acquire damage information. Then the last step is to calculate the various damage losses. Therefore, the system functions contain damage information extraction module, loss assessment module, information service module and database module. Damage information extraction function includes modules of sample, training sample's acquisition, and damage information extraction and so on. The losses assessment function contains the modules of counting building's damage rate; computation area to all building's, and estimate for other property and lives'losses. The information service function realizes the data's display, maintenance, spatial analysis. Almost all of the functions above need assessment data to support. The system database stores and manages raster, vector and other type of data. According to these data's purpose, the system database is understood to makes up of images database, GIS database, sample images database and estimate database, which is set up on ORACLE 10g and ArcSDE9. The system is tested by an instance, and the result indicates the assessment based on RS is reasonable and useful for emergency.