基于伪中值滤波和小波变换的红外图像增强方法

针对红外图像对比度差、信噪比低的特点,提出一种基于伪中值滤波和小波变换的红外弱小目标增强的算法。图像首先经过伪中值滤波滤除部分噪声,然后进行小波变换得到小波系数,对小于阈值的系数利用基于临近系数保留法进行滤波,防止小目标误判为噪声被滤除,对于大于阈值的系数进行非线性增强,最后重构得到去噪后的增强图像。实验结果表明,本文算法有效地滤除了噪声,提高了图像的对比度,更符合人眼视觉特性。

Infrared image enhancement based on pseudo median filter and wavelet transformation

For infrared image with low contrast and low signal-to-noise ratio,an infrared image enhancement method based on pseudo median filter and wavelet transformation is presented.Firstly,pseudo median filter is used for de-noising,then the wavelet transformation is adopted to decompose the image.To prevent the small target to be filtered as noise,the coefficient below threshold is filted by algorithm based on neighbor coefficient keeping.The coefficient above threshold is enhanced by nolinear method.Finally,the inverse wavelet transform is applied to synthesis image which can obtain the enhanced image.A group of experiment results show that the presented algorithm reduces the noise and improves the image contrast.The result image fits visual sense characteristic.