一种级联的激光主动成像图像融合降噪算法

针对激光主动成像图像混合噪声的特点,提出了一种投票中值滤波和整数提升小波级联的融合降噪算法。首先对激光图像进行噪声像素点检测,区分噪声点与非噪声点；而后采用投票中值滤波对噪声点进行处理,抑制脉冲噪声；然后采用整数提升小波变换对图像进行Bayes自适应阈值去噪,抑制散斑噪声；最后通过逆变换得到去噪图像。通过实验比对结果表明,该方法在具有良好的去噪、边缘保持性能的同时,还具有较为理想的实时性。

Cascaded image fusion de-noising algorithm for laser active imaging

Aiming at complex noise of laser active imaging,an image fusion de-noising algorithm based on the voting median filter and integer lifting-wavelet transform was proposed.Firstly,the noise pixels of laser image were detected to distinguish noise points and non-noise points.Then,voting median filter was used to deal with noise points and suppress impulse noise.Besides,the noise image was decomposed with integer lifting-wavelet transform,and wavelet coefficients were processed with Bayes adaptive threshold method.Finally,the de-noised image was obtained by inverse transform.By analyzing the experimental results,this method has good de-noising performance and edge retention,as well as ideal real-time.