遥感成像噪声分析及基于PCNN的滤除方法

Imaging noise analysis and a filter based on PCNN

遥感图像噪声的来源比较复杂,在遥感成像和图像传输过程中都有噪声的影响,所以,遥感成像噪声分析及滤除在遥感图像处理中占据非常重要的地位.CCD(电荷耦合器件)噪声是成像过程中最主要的噪声,它主要有暗流、零响应偏移和响应不一致等特性.滤除CCD噪声中暗流的直流分量以及消除零响应偏移和响应不一致后,可以认为所剩CCD噪声为高斯噪声和椒盐噪声的叠加.针对CCD噪声的特点和脉冲耦合神经网络(PCNN)中神经元间捕获特性的滤波作用,提出基于PCNN的CCD噪声滤波方法,并进行了仿真计算,验证了该方法的有效性.

There are many kinds of complex noises in imaging and data transferring from the remote sensing satellite. Imaging noise analysis and filter are very important for us to get a more precise image. CCD (Charge Coupled Device) is the most popular device used in the satellite to obtain image and the noises caused by CCD effect the image greatly. The main noises of CCD are dark current, response bias and response nonuniformity. After filtering the above noises the residual CCD noises can be regarded as Gauss noise and pepper-salt noise. According to the pulse coupled neural networks' (PCNN) work principle and the feathers of CCD noises, a filter based on PCNN was produced. And a simulation computation result was given to prove its feasibility and efficiency.