基于HJ-1卫星的重庆市都市功能区气溶胶光学厚度反演研究

发布时间：2018-01-22 12:03

本文关键词： 气溶胶 HJ-1A/B 暗像元法深蓝算法 PM10　出处：《重庆师范大学》2016年硕士论文　论文类型：学位论文

【摘要】：随着我国经济的高速发展,城市化进程加快,大气污染问题日益严重。近年来,大气污染颗粒已经成为城市污染的首要污染物,监测大气气溶胶已逐渐成为日常监测业务,但是基于点的监测不仅不能反映区域大气污染的空间分布状况,而且监测设备费用昂贵,而遥感技术的特点使它快速逐渐成为城市大气污染监测的手段之一。目前对于气溶胶光学厚度的反演方法应用较广的是暗像元法(DDV),但是对于高反射区域的城市地表适用性并不高,由于深蓝算法适用高反射地表,本文研究使用HJ-1A/B的高分辨率数据结合暗像元法和深蓝算法反演重庆市都市功能区的大气气溶胶光学厚度。首先,详述暗像元法和深蓝算法的原理与区别;其次,分别应用暗像元法和深蓝算法以及暗像元结合深蓝算法进行研究区上空的AOD进行反演,判别最适宜研究区反演AOD的研究方法;最后,将反演得到的AOD与实测的PM10进行相关分析,分析研究区PM10空间分布状况以及趋势变化。(1)确定暗像元的判别方法,采用的是CCD相机数据加入绿光波段的EVI进行暗像元判别。(2)应用DDV反演研究区上空AOD的时候,建筑密集的区域是未完全进行反演,无法满足进行城市气溶胶反演的要求;而应用深蓝算法虽然能够全部区域进行反演,且与城市人口、建筑分布有着较高的一致性,但是反演的AOD值普遍偏高,这与实际的状态又不相符;用DDV-深蓝算法结合方法进行反演,结果更加符合城市气溶胶反演的要求与实际分布情况。(3)将反演的结果与CE-318实测数据进行误差分析,绝对误差的最大值也在允许的范围内,证明文中所用反演方法具有可行性。(4)从春、夏、秋、冬四个季节的气溶胶光学厚度值进行分析。结果表明春季特别是初春时,气溶胶光学厚度比其他季节较高;夏季,季风带走了弥漫在空气中的气溶胶颗粒以及其他大气污染的颗粒物,使得气溶胶的移动有了明显的路径变化;秋季,由于重庆的地理环境的因素,开始在上空聚集大量的污染物粒子,能见度降低;迈入冬季时,城市上空常常弥漫着烟雾,能见度不高,空气的质量很差。(5)将反演结果进行气溶胶标高的订正与进行了湿度订正后的PM10进行线性分析,R2为0.505,相关性良好,模拟两者之间的方程式,进行PM10的反演,得到了PM10的空间分布特征。
[Abstract]:With China's rapid economic development, city urbanization, air pollution has become a serious problem. In recent years, the primary pollutant particulate air pollution has become a city of pollution, atmospheric aerosol monitoring has become a routine monitoring business, but the distribution of the monitoring point can not reflect regional air pollution based on the spatial and monitoring equipment is expensive however, the characteristics of remote sensing technology makes it become one of the rapid monitoring of atmospheric pollution in the city. The inversion method of aerosol optical thickness was widely applied in dark pixel method (DDV), but for the city surface applicability of high reflection region is not high, because the deep blue algorithm for high reflection surface, this paper studies high resolution use the data of HJ-1A/B combined with aerosol optical dark pixel method and the inversion algorithm of blue Chongqing city function region thickness. First,. The principle and the difference between dark pixel and deep blue algorithm; secondly, respectively using dark pixel and dark blue with dark blue dark pixel algorithm and algorithm research over the region AOD inversion, the most suitable research methods of distinguishing the inversion of AOD; finally, AOD and PM10 were measured by inversion analysis, analysis the study area of PM10 spatial distribution and change trend. (1) to determine the method of distinguishing dark pixels, using the CCD camera data into the green band EVI dark pixel discrimination. (2) when using DDV over AOD inversion in the study area, building dense area is not completely unable to meet the inversion, the city of aerosol inversion requirements; and the application of deep blue algorithm can all regional inversion and city population distribution, the building has a higher consistency, but the inversion of AOD value is generally high, and the actual The state is not consistent; inversion using DDV- algorithm combined with blue method, results more in line with the city aerosol retrieval requirements and actual distribution. (3) the inversion results and the CE-318 data error analysis, the maximum absolute error is within the allowable range that the inversion method is feasible. (4) from the spring, summer, autumn, winter four season of aerosol optical thickness values were analyzed. The results show that the spring is spring, the aerosol optical thickness higher than other seasons; summer monsoon, particles away in the air aerosol particles and other air pollution, the aerosol is mobile the path change obviously; autumn, due to the geographical environment of Chongqing, began to gather a large number of particle pollutants, over reduced visibility; into the winter, often filled with smoke over the city, can The visibility is not high, and the air quality is very poor. (5) the correction of aerosol elevation is corrected by the inversion result, and the PM10 of the corrected humidity is linearly analyzed. R2 is 0.505, and the correlation is good. The equation between the two is simulated, and PM10 is retrieved, and the spatial distribution characteristics of PM10 are obtained.

【学位授予单位】：重庆师范大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：X87;X513

【相似文献】