摘    要目前我国在环境问题上面临着重大的问题，其中大气污染中的雾霾问题是最为突出的。它严重影响着我国城市和区域的可持续发展以及人民群众的身体健康，而产生这一问题的主要原因是由于我国城市化进程加快，各地人群涌入大城市。同时工业化进程也在迅速发展，过量的工业气体排放，拥挤的交通运输以及大量的煤炭燃烧等各种人为因素造成雾霾现象严重。所以全面掌握大气污染的区域分布状况和其变化规律对环境治理和规划极为重要。并且随着定量遥感技术的发展，利用遥感数据进行PM2。5浓度监测有着重要意义。90003

本文将利用Terra-MODIS卫星的气溶胶数据，选取2014年到2015年的晴天影像建立样本，该数据对外免费开放，具有很好的经济适用性。与此同时，还获取了相对应时间段的杭州市各监测点的瞬时PM2。5浓度监测平均值，与之相对应时间的气象因子数据。首先对PM2。5实测数据和气象数据进行基本的趋势变化分析。然后再对获取的MODIS 影像进行气溶胶反演处理，我们将利用现如今理论体系较为成熟的暗像元算法和6S辐射传输模式，对在浙江省的杭州市区域范围内的气溶胶光学厚度进行反演，并选取的反演好的结果数据（主要是2014-2015年的其中五个月）进行基本的个例分析。最后利用空间统计分析原理对杭州市PM2。5浓度的整体特征分别在空间和时间维度进行分析，并选取气溶胶光学厚度反演结果中质量最好的那一个结果，将气溶胶光学厚度值与监测站得到的PM2。5浓度值做多种类别的回归性建模分析。这里主要是建立以线性函数、二次函数、三次函数、指数函数、对数函数和幂函数为基础的回归模型，通过对回归模型拟合优度R²和模型检验精度的对比选择出最佳拟合模型。以此来讨论并说明哪一种模型是最适合分析运用于大气气溶胶光学厚度和PM2。5之间的关系。

毕业论文关键词：PM2。5； 气溶胶光学厚度； 暗像元法； MODIS； 回归分析； 相关分析

At present, China is facing major problems in environmental issues where the haze problem of air pollution is the most prominent。It seriously affects the sustainable development of China's cities and regions and the health of the people,The main reason for this problem is due to the accelerated process of urbanization in China and  the influx of people around the big cities。At the same time the process of industrialization is also rapid development, excessive industrial gas emissions, crowded transport and a large number of coal burning and other human factors caused by foggy serious。Complete control of air pollution in the area of distribution and its variation is extremely important for environmental management and planning,and with the development of quantitative remote sensing technology, the use of remote sensing data for PM2。5 concentration monitoring is of great significance。

This article will use the Terra-MODIS satellite aerosol data,and select a sunny day image from 2014 to 2015,because of the data is open for free and has good economic applicability。Using the modern system theory is more mature dark pixel algorithm and 6S radiation transmission mode,The aerosol optical thickness in the area of Hangzhou is inversed and according to the season for a case study。At the same time,I also obtained the corresponding time period Hangzhou monitoring point instant PM2。5 concentration monitoring value。The spatial and temporal dimensions of the PM2。5 concentration in Hangzhou were analyzed by using the principle of spatial statistical analysis。Finally, I select the best part of the aerosol optical thickness inversion results。The correlation between aerosol optical thickness and PM2。5 concentration obtained from the monitoring station was analyzed and regression analysis。Here is mainly to establish in a linear quadratic function ,three function, index function regression model on the basis of logarithmic function and power function。The best fitting model is selected by comparing the regression model goodness R² and the model test precision。