基于偏微分方程和机器视觉的模糊激光图像增强研究  

Research on fuzzy laser image enhancement based on partial differential equations and machine vision

作　　者：秦小娜[1] 刘文华[1] QIN Xiaona;LIU Wehua(Jiangxi University of Technology,Nanchang 330098,China)

机构地区：[1]江西科技学院,南昌330098

出　　处：《激光杂志》2023年第9期98-102,共5页Laser Journal

基　　金：国家自然科学基金项目(No.71663027);江西省教育科学“十四五”规划(No.21YB226)。

摘　　要：激光图像在采集与传输中受环境因素影响存在多种噪声信号,容易变得模糊,信噪比较低。为了提高模糊激光图像的增强效果,提升信噪比,提出基于偏微分方程和机器视觉的模糊激光图像增强方法研究。采用紧小波变换原理分析模糊现象原因;利用偏微分方程Perona-Malik模型去除图像噪声;基于机器视觉模型划分图像视觉显著区域和背景区域,结合Retinex理论从全局角度实现模糊激光图像的增强。实验结果表明:应用提出方法后激光图像信噪比与对比度均得到了显著提升,最大值分别为96 dB与0.94,提高了图像增强效果。In the process of laser image acquisition and transmission,there are many kinds of noise signals affected by environmental factors,which are easy to become blurred and the signal-to-noise ratio is low.In order to improve the enhancement effect for fuzzy laser image and enhance the signal-to-noise ratio,a fuzzy laser image enhancement method based on partial differential equation and machine vision is proposed.The principle of tight wavelet transform was used to analyze the causes of fuzziness.The partial differential equation and Perona Malik model are used to remove image noise;Based on the machine vision model,the salient area and the background area of the image are divided,and the enhancement of the fuzzy laser image is realized from the global perspective by combining Retinex theory.The experimental results show that the signal-to-noise ratio and contrast of the laser image are significantly improved after the proposed method is applied,and the maximum values are 96 dB and 0.94 respectively,which improves the image enhancement effect.

关 键 词：模糊图像 机器视觉 激光图像 偏微分方程 图像去噪 图像增强 

分 类 号：TN911.73[电子电信—通信与信息系统]