摘要
为满足高精度的激光测量图像质量标准、有效处理图像中的复杂噪声,提出基于模式识别的激光图像中噪声分类和去噪方法。根据长短记忆网络确定噪声的分布图像,通过NSCT分解算法对获取的噪声图像进行分解和重构,利用加权引导滤波算法处理滤波重构后的图像,输出去噪后的激光图像。不同环境下激光图像测试结果显示:该方法提升了图像整体清晰度和质量,去噪后图像的峰值信噪比均在30 dB以上,图像方差均在0.015以下,为激光图像后续处理提供可靠依据。
To meet high-precision laser measurement image quality standards and effectively handle complex noise in images,a pattern recognition based method for noise classification and denoising in laser images is proposed.Determine the distribution image of noise based on long short-term memory network,decompose and reconstruct the obtained noise image through NSCT decomposition algorithm,process the filtered reconstructed image using weighted guided filtering algorithm,and output the denoised laser image.The results of laser image testing in different environments show that this method improves the overall clarity and quality of the image.After denoising,the peak signal-to-noise ratio of the image is above 30 dB,and the image variance is below 0.015,providing a reliable basis for subsequent processing of laser images.
作者
孙珊珊
王鹏
SUN Shanshan;WANG Peng(School of Information Engineering,Suihua University,Suihua Heilongjiang 152000,China)
出处
《激光杂志》
北大核心
2025年第12期181-186,共6页
Laser Journal
基金
黑龙江省省属本科高校基本科研项目(No.YWF10236240128)。
关键词
模式识别
激光图像
噪声分类
去噪处理
NSCT分解
加权引导滤波
pattern recognition
laser image
noise classification
denoising processing
NSCT decomposition
weighted guided filtering
