基于改进Retinex算法的海底地貌图像增强方法  

Image Enhancement Method of Seabed Landform Based on Improved Retinex Algorithm

作　　者：常列 常淼 CHANG Lie;CHANG Miao(Guangzhou Hydrographic Center,Guangzhou 510000 China)

机构地区：[1]广州海事测绘中心,广东广州510000

出　　处：《自动化技术与应用》2024年第9期77-80,共4页Techniques of Automation and Applications

摘　　要：为获取更清晰的海底地貌图像,研究基于改进Retinex算法的海底地貌图像增强方法。使用高斯低通滤波器进行低频入射分量的卷积操作,计算图像带色彩恢复输出值;构造一般表现形式,得到像素区域内对数修正的输出值;基于多尺度Retinex算法对色彩对比度进行增强处理;计算亮度值以及图像的入射分量,对海底地貌图像进行光照补偿;改进Retinex算法,实现海底地貌图像增强。由对比处理图像结果可知,所提算法的增强图像亮度更均匀、色彩更饱满且细节纹理更清晰。在客观参数对比中,所提算法对比度、清晰度以及信息熵也高于其他图像,有效提高了图像清晰度。In order to obtain a clearer image of seabed landforms,an enhanced method of seabed landform images based on the improved Retinex algorithm is studied.It uses Gaussian low-pass filter to perform rolling operation of low-frequency incident components,and calculates the output value of image band color recovery,constructs a general representation to obtain the logarithmically corrected output value in the pixel area,enhances color contrast based on multi-scale Retinex algorithm.It calculates the brightness value and the incident component of the image,and performs illumination compensation on the seabed landform image,improves the Retinex algorithm to realize the enhancement of the seabed landform image.It can be seen from the comparison processing image results that the enhanced image brightness of this algorithm is more uniform,the color is fuller,and the detail texture is clearer.In the objective parameter comparison,the contrast,clarity and information entropy of this algorithm are also higher than other images,which effectively improves the image clarity.

关 键 词：图像处理 改进Retinex算法 海底地貌图像 图像纹理细节 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]