目的 研究消除线阵CCD 在采集图像过程中引入的噪声。 方法 根据印刷品检测系统中常引入高斯噪声和椒盐噪声结果， 分别应用小波变换算法和高斯-中值混合算法进行去噪处理。 结果 通过与不同方法的对比分析， 表明该 2种方法可行。 结论 小波变换对高斯噪声图像能达到很理想的去噪效果， 高斯-中值混合算法实现了在去噪的情况下还保持图像本身的细节部分， 对椒盐噪声去噪效果明显。

Objective To study the problem of eliminating the noises introduced during image collection process of linear array CCD. Methods Through experiments, it was found that Gaussian noise and salt pepper noise were usually introduced into the detection system of prints. Wavelet transforms and Gaussian-median mixed algorithm were used in the denoising processing. Results Through comparison with different methods, the feasibility of these two methods were proved. Conclusion Wavelet transforms was good in removing the Gaussian noise, while Gaussian-median mixed algorithm achieved good results in eliminating slat pepper noise because it could keep the details of the images during denoising.