基于双重Gamma校正的秧苗图像增强算法  

Image enhancement algorithm for seedling image with dual gamma correction

作　　者：廖娟 陈民慧 汪鹞 邹禹 张顺[1] 张培江[2] 朱德泉[1] LIAO Juan;CHEN Min-hui;WANG Yao;ZOU Yu;ZHANG Shun;ZHANG Pei-jiang;ZHU De-quan(School of Engineering,Anhui Agricultural University,Hefei 230036,China;Rice Research Institute,Anhui Academy of Agricultural Sciences,Hefei 230031,China)

机构地区：[1]安徽农业大学工学院,安徽合肥230036 [2]安徽省农业科学院水稻研究所,安徽合肥230031

出　　处：《江苏农业学报》2020年第6期1411-1418,共8页Jiangsu Journal of Agricultural Sciences

基　　金：国家重点研发计划项目(2018YFD0700304);安徽省重点研发计划项目(202004a06020016);安徽省科技重大专项(18030701204)。

摘　　要：为了提高复杂光照条件下水田秧苗图像的视觉效果,提出1种基于双重Gamma校正的秧苗图像增强算法。将原始RGB[红(R)-绿(G)-蓝(B)]图像转换成HSV[色调(H)-饱和度(S)-亮度(V)]颜色空间图像,对V分量进行亮度区域划分;通过快速引导滤波法提取秧苗图像的光照度分量,利用光照信息的分布特性自适应地设置Gamma控制参数,并构建2个自适应Gamma函数,实现对V分量图像亮度的独立校正;最后,对校正后的图像进行自适应融合,并结合H、S分量转换为RGB图像,进行色彩饱和度的恢复。结果表明,本研究算法能够实现不同光照条件下秧苗图像的自适应增强,丰富图像中的有用信息,保真原图像的色彩信息,有效改善图像的视觉效果,为后期的秧苗分割提供可靠的处理对象。To improve the visual effect of seedling images under complex illumination conditions,an image enhancement algorithm based on dual gamma correction was proposed.The original RGB image was firstly converted into HSV color space image,and the V component was extracted and partitioned into dark and bright regions.The illumination component of seedling image was obtained by the fast guided filtering method.The gamma control parameters were dynamically adjusted by the local distribution characteristics of illumination,and two adaptive gamma functions were constructed to correct the luminance of V-component image.Then,the corrected images were fused adaptively.Combining with H and S components,the fused V image was used to obtain the RGB image.Finally,the color saturation restoration was carried out for RGB image to solve the color deviation.Experimental results show that the algorithm can achieve adaptive enhancement of seedling images under different lighting conditions,enrich the useful information in the image,maintain the color information of the original image,effectively improve the visual effect of the image,and provide reliable processing objects for seedling segmentation.

关 键 词：秧苗图像 图像增强 自适应Gamma校正 色彩饱和度修复 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]