中文摘要：

灰霾环境对遥感图像具有非常严重的影响。主要分析了太阳光在灰霾环境中的传输方法。根据太阳光的传输方式,把灰霾层看作是一个不均匀的半透明-透明层。然后通过SMACC算法提取灰霾的端元信息,并利用最小二乘法反演去除灰霾后的遥感图像。本文所述的线性光谱分解去霾方法对遥感图像中气溶胶浓度低/高的灰霾的去除效果都比较好,而且较大限度地保留了图像的光谱信息。最后把本文方法和大气校

英文摘要：

Haze has great influence on remote sensing images.The transmission of sunhght in haze environment is analyzed mainly.Firstly,according to the transmission of sunlight,the haze is considered as a non-uniform translucent-transparent layer.Then,the end-member of the haze is extracted by using the SMACC algorithm and the haze in a remote sensing image is removed by using a least-square method. The haze removal method using hnear spectral decomposition is very effective for the removal of the haze with either high or low aerosol concentration in remote sensing images and can retain the spectral information in remote sensing images maximally.Finally,the method is compared with the atmospheric correction method.The result shows that this linear spectral decomposition method is better for the removal of haze.