一种改进的图像阴影去除方法

doi:10.16180/j.cnki.issn1007-7820.2021.04.007

电子科技 ›› 2021, Vol. 34 ›› Issue (4): 41-46.doi: 10.16180/j.cnki.issn1007-7820.2021.04.007

一种改进的图像阴影去除方法

胡晨,胡德敏,胡钰媛,褚成伟

上海理工大学光电信息与计算机工程学院,上海200093

收稿日期:2019-12-30 出版日期:2021-04-15 发布日期:2021-04-16
作者简介:胡晨(1993-),男,硕士研究生。研究方向:图像处理、深度学习。|胡德敏(1963-),男,博士,副教授。研究方向:计算机网络、分布式计算、云计算。|胡钰媛(1995-),女,硕士研究生。研究方向:图像处理。|褚成伟(1995-),男,硕士研究生。研究方向:自然语言处理、深度学习。
基金资助:
国家自然科学基金(61170277);国家自然科学基金(61472256);上海市教委科研创新重点项目(12ZZ17);上海市一流学科建设项目(S1201YLXK)

An Improved Image Shadow Removal Method

HU Chen,HU Demin,HU Yuyuan,CHU Chengwei

School of Optocal-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China

Received:2019-12-30 Online:2021-04-15 Published:2021-04-16
Supported by:
National Natural Science Foundation of China(61170277);Fund:National Natural Science Foundation of China(61472256);Key Project of Scientific Research and Innovation of Shanghai Municipal Education Commission(12ZZ17);First Class Construction Project of Shanghai(S1201YLXK)

摘要/Abstract

摘要：

针对阴影去除后存在光照不饱和与纹理不一致等问题,文中提出一种基于Retinex的阴影去除方法。该方法依据Retinex理论中入射光与反射光的关系,重新估算入射光图,用RGB三通道中像素的平均值赋值给目标像素点,并用梯度下降方法进行改进。通过调节自适应的权重,改善了阴影去除后的光照不饱和与纹理不一致问题。在UIUC数据集上的实验表明,实验结果的均方根误差比降低到0.4106,证明了文中方法的有效性。

关键词: HSV, 归一化饱和差分指数, 最大类间方差法, Retinex, 光照不饱和, 边缘平滑, 纹理增强, 梯度下降法

Abstract:

In view of the problems of illumination unsaturation and texture inconsistency after shadow removal, a shadow removal method based on Retinex is proposed in this study. Based on the relationship between incident light and reflected light in Retinex theory, this method reevaluates the incident light map, assignes the average value of pixels in the three RGB channels to the target pixel point, and improves the gradient descent method. By adjusting the adaptive weight, the problem of illumination unsaturation and texture inconsistency after shadow removal is improved. The experiments on the UIUC data set prove that the root mean square error ratio of the experimental results is reduced to 0.4106, which verifies the effectiveness of the proposed method.

Key words: HSV, NSVDI, Nobuyuki Otsu, Retinex, light unsaturated, edege smooth, texture enhancement, gradient descent method

中图分类号:

TP391.41

胡晨,胡德敏,胡钰媛,褚成伟. 一种改进的图像阴影去除方法[J]. 电子科技, 2021, 34(4): 41-46.

HU Chen,HU Demin,HU Yuyuan,CHU Chengwei. An Improved Image Shadow Removal Method[J]. Electronic Science and Technology, 2021, 34(4): 41-46.

图/表 7

图1

表1

图2

图3

表2

表3

图4

参考文献 18

[1]	Xu M L, Zhu J J, Pei L , et al. Learning based shadow recognition and removal from monochromatc natural images[J]. IEEE Transactions on Image Process- ig, 2017,26(12):5811-5824.
[2]	Tiemaxiaosu. 阴影处理[EB/OL].( 2016 -06-23)[2019-09-15] https://blog.csdn.net/tiemaxiaosu/article/details/5164871.
[3]	Zhu J, Kegan S G G, Marshall S Z. Learning to recognize shadow in monochromatic images[C]. San Francisco: IEEE Conference on Computer Vision and Pattern Recognition, 2010.
[4]	Gryka M, Terry M, Brostow G J. Learning to remove soft shadows[J]. ACM Transactions on Graphics, 2015,34(5):1-15.
[5]	Khan S H, Bennamoun M, Sohel F , et al. Automatic shad- ow detection and removal from a single image[J]. IEEE Transactions on Pattern Anal Mach Intell, 2016,38(3):431-446.
[6]	Guo R Q, Dai Q Y, Derek H. Paired regions for shadow detection and removal[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013,35(12):2956-2967. doi: 10.1109/TPAMI.2012.214 pmid: 24136433
[7]	梁磊, 刘怀, 董春燕 , 等. 结合亮度比值梯度与颜色特征的阴影去除算法[J]. 小型小微计算机系统, 2019,40(4):802-807.
	Liang Lei, Liu Huai, Dong Chunyan , et al. Shadow removal algorithms combining brightness ratio gradient and color feature[J]. Journal of Chinese Computer Systems, 2019,40(4):802-807.
[8]	Zhang L, Zhang Q, Xiao C. Shadow remover:image shadow removal based on illumination recovering optimization[J]. IEEE Transactions on Image Processing, 2015,24(11):1-13.
[9]	帅滔, 张洪艳. 基于新型阴影指标的遥感影响阴影检测方法[J]. 电子科技, 2016,29(2):92-95.
	Shuai Tao, Zhang Hongyan. A shadow detection method in high-resolution remote sensing imagery based on novel shadow index[J]. Electronic Science and Technology,206, 29(2):92-95.
[10]	Saritha Murali, Govindan V K. Shadow detection and removal from a single image using LAB color space[J]. Cybernetics and Information Technologies, 2013,13(1):95-103.
[11]	Yana I S. Shadow detection and removal using a shadow formation model[C]. Lviv:IEEE First International Conference on Data Stream Mining & Processing, 2016.
[12]	Victor J D T. A comparative study on shadow compesation of color aerial images in invariant color models[J]. IEEE Transactions on Geoscience and Remote Sensing, 2006,44(7):1661-1671.
[13]	鸭子船长. 颜色空间RGB与HSV(HSL)的转换[EB/OL].( 2015 -10-27)[2019-9-15]https://www.cnblogs.com/zl1991/p/4913450.html.
[14]	Krishna K S, Kirat P, Nigam M J. Shadow dete- ction and removal from remote sensing images using NDI and morphological operators[J]. International Journal of Computer Applications, 2012,42(10):37-40.
[15]	Azevedo S C, Silva E A, Pedrosa M M. Shadow detection improvement using spectral indices and morphological operators in urban areas in high resolution images[C]. Berlin:The Thirty-sixth International Symposium on Remote Sensing of Environmental, 2015.
[16]	Imageshop. 带色彩恢复的多尺度视网膜增强算法(MSRCR)的原理、实现及应用[EB/OB].( 2013 -04-17) [2019-08-26]https://www.cnblogs.com/Imageshop/archive/2013/04/17/3026881.html.
[17]	Land E H. The retinex theory of color vision[J]. Scientific American, 1977,237(6):108-128. doi: 10.1038/scientificamerican1277-108 pmid: 929159
[18]	Han G, Darren C. Image and vision computing[J]. Image and Vision Computing, 2017,62(4):19-27.

参数	数值
处理器	第八代智能英特尔酷睿 i5-8300H
内存	8 GB
操作系统	Windows 10 64位
编程软件	Python3.6+OpenCV3(纯CPU模式)

方法	BER
文献[1]	11.003 2
文献[6]	12.678 5
文献[10]	12.293 6
本文方法	10.358 1

方法	ER(image)	ER(shadow)
文献[1]	0.621 1	0.439 2
文献[6]	0.634 0	0.526 8
文献[10]	0.621 6	0.447 8
本文方法	0.601 5	0.410 6

一种改进的图像阴影去除方法

An Improved Image Shadow Removal Method

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 7

参考文献 18

相关文章 14

Metrics

本文评价

推荐阅读 6

[1]	杨凌霄,王振营,刘群坡,王高伟,杨彦超. 基于H-S直方图的苹果着色率检测算法研究[J]. 电子科技, 2021, 34(5): 13-17.
[2]	张翔,史志才,陈良. 基于SWA优化级联网络的表情识别方法[J]. 电子科技, 2020, 33(9): 16-20.
[3]	马旭东,袁锐波. 基于图像检测的硬币面值识别方法[J]. 电子科技, 2020, 33(2): 43-47.
[4]	王炜,胡一鸣,石强. 梯度下降法在水库调度函数优化中的应用[J]. 电子科技, 2020, 33(1): 68-72.
[5]	胡德敏,程普芳. 一种改进的LeNet网络[J]. 电子科技, 2019, 32(3): 53-57.
[6]	杨俊,魏静萱. 梯度策略自适应差分进化算法[J]. , 2016, 29(1): 25-.
[7]	郁浩,张荣福,程金光,郭世平. 基于图像处理的钢包标牌识别方法[J]. , 2016, 29(1): 94-.
[8]	余婷. 基于结构相似度的图像去噪新方法[J]. , 2015, 28(3): 1-.
[9]	赵日成. 基于车底阴影的车前障碍物检测[J]. , 2015, 28(3): 15-.
[10]	王毅,王创新,卢进,盛文正. 基于HSV模型和改进AdaBoost算法的车牌检测[J]. , 2015, 28(2): 107-.
[11]	马学智, 吕宁. 一种红外焦平面阵列的盲元检测快速算法[J]. , 2013, 26(10): 88-.
[12]	尹红然, 俞浩, 秦清锋. 一种基于Retinex理论的彩色图像增强算法[J]. , 2012, 25(3): 21-.
[13]	郑瀚, 韦文斌, 齐子城. 机器视觉检测胶囊图像的预处理研究[J]. , 2012, 25(12): 133-.
[14]	常蓬勃, 李丹红. 粗糙集方法在红外图像增强中的应用[J]. , 2010, 23(6): 43-.