基于同态滤波的遥感图像云雾去除模型研究

Cloud Removal from Remote Sensing Image Using Homomorphic Filtering

在遥感卫星成像过程中,云雾作为一种噪声时常导致传感器接收到的地物辐射与真实值之间存在差距,是影像降质的主要原因。本文以均匀薄云雾覆盖下的SPOT-5遥感图像为研究对象,基于其频域中云雾的低频特征,建立了同态滤波云雾去除模型,并从主观评价及客观分析两方面对比评估了该模型的应用效果。结果表明,基于指数高通滤波器构建的同态滤波去云雾模型能够有效地提升云雾遮盖下的影像质量。

In the process of remote sensing satellite imaging,cloud as a kind of noise often leads to the gap between the ground radiation received by the sensor and the true value,which is the main reason for image degradation.In this paper,the SPOT-5 remote sensing image under uniform thin cloud cover is taken as the research object.Based on the low frequency characteristics of the cloud in the frequency domain,the homomorphic filtering cloud removal model is established,and the model is compared and evaluated from subjective evaluation and objective analysis.Application effect.The results show that the homomorphic filtering de-clouding model based on exponential high-pass filter can effectively improve the image quality under cloud cover.