【作者】 刘勇；

【导师】 查勇；

【作者基本信息】 南京师范大学 ， 地图学与地理信息系统， 2007， 硕士

【摘要】 国民经济的快速发展，城市规模的不断扩大，城市各类工程的不断开展，机动车的不断增加，给城市空气质量带来了严峻的挑战。实践证明遥感在城市空气质量监测中发挥了重要的作用。本文利用MODIS遥感反演的气溶胶光学厚度数据和地面监测的空气污染指数数据，通过两者之间的相关分析，试图建立相应的关系模型。此外由于空气污染状况以及气溶胶光学厚度也与一定的气象条件有关，因此引进气象因子参与关系讨论，并建立相应模型，以找出空气污染指数与气溶胶光学厚度之间的较好模型，从而实现利用遥感手段监测城市空气污染状况的目的。通过研究得到如下结论：1、全年、季节空气污染指数和气溶胶光学厚度建立回归模型，比较各模型精度，得出夏季线性模型和秋季线性模型预测精度较理想，夏季相关系数达到0.853，秋季相关系数达到0.838。2、讨论常规的气象观测数据，选取主导气象因子为风速和气压，考虑两气象因子的作用，对全年和季节模型进行重建，发现有主导气象因子参与的多元回归模型预测精度普遍比没有气象因子参与的一元回归模型高。3、依据主导气象因子风速和气压将样本数据分组，各组空气污染指数和气溶胶光学厚度建立回归模型，并讨论其精度，发现气压大于1020hpa且风速小于1m／s时，空气污染指数与气溶胶光学厚度之间的模型预测精度最理想，预测精度达到95.4％。4、对季节样本数据在气象因子分级后的情况进行讨论，发现夏季和秋季使用气象因子分级模型预测API值比这两个季节本身模型预测效果要好，精度平均提高2个百分点，春季和冬季效果不是很明显。5、通过本文的研究，发现在考虑气象因子的情况下，探讨空气污染指数和气溶胶光学厚度的关系，会得到更好的效果，更能真实地反映地面空气的污染状况。更多还原

【Abstract】 The fast development of national economy, the unceasing expansion of city scale, the unceasing development of all kinds of urban projects, as well as the unceasing development of motor vehicles, all of which had brought the stern challenge to the city air quality. The practice proved that remote sensing had played an important role in urban air quality monitor. This paper used the MODIS Aerosol Optical Thickness data and the ground monitor Air Pollution Index data, attempting to establish the corresponding relational model by the correlation analysis between API and AOT.In addition, the air pollution condition or the aerosol optical thickness was related to certain meteorological condition, so the meteorological factors must be discussed when studying the relationship between the API and AOT. Then established the corresponding model to find out the good model of API and AOT, thus realized the purposes of using remote sensing to monitor the status of urban air pollution. The main conclusions of this paper were drew as follows:1. Through regression analysis to API and AOT of season or year to compare each model’s precision, found that the predicted accuracy of summer and autumn model were good.2. Through discussing the conventional meteorological observation data, the wind speed and the barometric pressure were selected as the main meteorological factor. Then further considered the above two meteorological factors in the model, which were reconstructed by the whole year and the season data, it was found that the precision of the model involved the meteorological factor was higher than which no meteorological factors involved.3. The sample data were grouped according to the main meteorological factor such as wind speed and barometric pressure. Through regression analysis to each group’s API and AOT, and discussed its accuracy. It was found that the model’s precision between API and AOT was the highest and the forecast accuracy reached 95.4%, when the barometric pressure was more than 1020hpa and the wind speed was less than 1m/s.4. Through discussing seasonal sample data after the meteorological factor grading, found that the forecast results of the predictive API by meteorological factor grading model was better than their own model in the summer and autumn, which the accuracy averagely increased of 2 percents points. The effect of spring and winter were not very significant.5. Through the research of this paper, we can find that discussing the relationship between API and AOT with the meteorological factors will obtain a better effect, and even really reflect the air pollution condition on the ground.更多还原