基于子窗口盒式滤波的MPPC微光成像增强算法

汤祥裕; 尹丽菊; 周辉; 邹国锋; 崔玉敏; 邓玉林; 张言华

doi:10.5768/JAO202243.0604008

基于子窗口盒式滤波的MPPC微光成像增强算法

doi: 10.5768/JAO202243.0604008

山东理工大学电气与电子工程学院，山东淄博 255049

基金项目: 国家自然科学基金（62101310）；山东自然科学基金（ZR2020MF127）

详细信息

作者简介:
汤祥裕（1996—），男，硕士研究生，主要从事微光图像处理方面的研究。E-mail：1275821162@qq.com

通讯作者:
尹丽菊（1972—），女，教授，博士，主要从事光电检测技术、成像探测与机器视觉方面的研究。E-mail：Ljyin72@163.com

中图分类号: TN223
计量
- 文章访问数: 107
- HTML全文浏览量: 60
- PDF下载量: 9
- 被引次数: 0
出版历程
- 收稿日期: 2022-05-05
- 修回日期: 2022-06-20
- 网络出版日期: 2022-09-24
- 刊出日期: 2022-11-14

MPPC low-level-light imaging enhancement algorithm based on sub-window box filtering

College of Electrical and Electronic Engineering, Shandong University of Technology, Zibo 255049, China

摘要

摘要: 针对多像素光子计数器(MPPC)进行微光成像时，图像受光照不足和噪声影响出现的图像亮度低、对比度差、边缘模糊等问题，提出一种基于子窗口盒式滤波的自适应微光图像处理算法。为了减少算法运行时间的同时突出图像的边缘细节信息，利用子窗口盒式滤波器对图像进行分层得到基础层和细节层；对基础层图像采用自适应阈值直方图均衡化拉伸对比度，细节层图像采用自适应增益控制方式进行增强；根据基础层图像中有效灰度值个数占总灰度的比值自适应确定融合系数，将基础层图像与细节层图像融合得到增强后图像。通过微光实验平台设置3组不同照度的微光环境进行实验仿真，验证了本文算法在保持边缘信息和增强细节方面获得了更好的效果。实验结果表明本文算法在标准差、信息熵、平均梯度等客观评价方面优于改进前算法，提升了微光图像的成像效果。
- 图像增强 /
- 微光环境 /
- 子窗口滤波 /
- 图像分层 /
- 自适应融合
Abstract: Aiming at the problems of low image brightness, poor contrast, and blurred edges caused by insufficient illumination and noise during low-level-light imaging for multi-pixel photon counter (MPPC), an adaptive low-level-light image processing algorithm based on sub-window box filtering was proposed. To reduce the algorithm running time while highlighting the edge detail information of the image, the sub-window box filter was used to layer the image to obtain the basic layer and detail layer. For the image of basic layer, the adaptive threshold histogram equalization was used to stretch the contrast, and the image of detail layer was enhanced by adaptive gain control method. The fusion coefficient was determined adaptively based on the ratio of the number of effective gray values to the total gray in the image of basic layer, and the image of basic layer was fused with the image of detail layer to obtain the enhanced image. Three sets of low-level-light environments with different illumination levels were set by the low-level-light experimental platform for experimental simulation, which verified that the algorithm obtained better results in maintaining edge information and enhancing details. Experimental results show that the proposed algorithm is superior to the previous algorithm in objective evaluation of standard deviation, information entropy, and average gradient, which improves the imaging effect of low-level-light image.
- image enhancement /
- low-level-light environment /
- sub-window filtering /
- image layering /
- adaptive fusion

HTML全文

图 1 不同类型支持区域的滤波特征

Fig. 1 Diagram of filtering characteristics in different types of support regions

下载: 全尺寸图片幻灯片

图 2 中心像素位于8个子窗口的位置

Fig. 2 Center pixel located in position of eight sub windows

下载: 全尺寸图片幻灯片

图 3 不同参数条件下的子窗口盒式滤波器

Fig. 3 Sub-window box filter under different parameter conditions

下载: 全尺寸图片幻灯片

图 4 子窗口盒式滤波器运算流程

Fig. 4 Flow chart of sub-window box filter operation

下载: 全尺寸图片幻灯片

图 5 引导滤波和本文滤波器保持边缘性能对比

Fig. 5 Performance comparison in edge preserving between guided filter and proposed filter

下载: 全尺寸图片幻灯片

图 6 本文微光图像增强算法流程

Fig. 6 Flow chart of proposed low-level-light image enhancement algorithm

下载: 全尺寸图片幻灯片

图 7 算法处理前后的微光图像直方图分布

Fig. 7 Histogram distribution of low-level-light image before and after algorithm processing

下载: 全尺寸图片幻灯片

图 8 累计直方图概率分布

Fig. 8 Cumulative histogram probability distribution

下载: 全尺寸图片幻灯片

图 9 直方图均衡化效果对比

Fig. 9 Effect comparison of histogram equalization in different scenes

下载: 全尺寸图片幻灯片

图 10 微光图像细节层增强

Fig. 10 Detail layer enhancement of low-level-light image

下载: 全尺寸图片幻灯片

图 11 实验平台实物图

Fig. 11 Physical drawing of experimental platform

下载: 全尺寸图片幻灯片

图 12 实验平台内部结构

Fig. 12 Internal structure of experimental platform

下载: 全尺寸图片幻灯片

图 13 不同照度下获取的原始图像

Fig. 13 Raw images obtained at different levels of illumination

下载: 全尺寸图片幻灯片

图 14 ${10^{ - 1}}\;{\rm{lx}}$微光环境下各算法增强图像效果对比

Fig. 14 Comparison of effect of image enhancement by algorithms under ${10^{ - 1}}\;{\rm{lx}}$ low-level-light condition

下载: 全尺寸图片幻灯片

图 15 ${10^{ - 3}}\;{\rm{lx}}$微光环境下各算法增强图像效果对比

Fig. 15 Comparison of effect of image enhancement by algorithms under ${10^{ - 3}}\;{\rm{lx}}$ low-level-light condition

下载: 全尺寸图片幻灯片

图 16 ${10^{ - 5}}\;{\rm{lx}}$微光环境下各算法增强图像效果对比

Fig. 16 Comparison of effect of image enhancement by algorithms under ${10^{ - 5}}\;{\rm{lx}}$ low-level-light condition

下载: 全尺寸图片幻灯片

表 1 ${10^{ - 1}}\;{\rm{lx}}$微光环境下的图像增强评价指标

Table 1 Image enhancement evaluation indexes under ${10^{ - 1}}\;{\rm{lx}}$ low-level-light environment

算法	标准差			信息熵			平均梯度
算法	A	B	C	A	B	C	A	B	C
原始图像	34.63	15.87	29.24	6.38	1.20	4.59	1.26	0.57	0.83
交叉子窗口算法	35.51	14.43	33.75	6.41	1.22	4.98	1.37	0.61	0.96
DDE算法	35.82	16.11	33.56	6.46	1.16	4.81	1.78	0.72	1.06
文献[4]算法	37.35	17.15	36.57	6.47	1.19	4.94	1.43	0.89	1.12
本文算法	39.58	20.56	40.36	6.49	0.77	5.19	2.27	1.15	1.46

下载: 导出CSV

表 2 ${10^{ - 3}}\;{\rm{lx}}$微光环境下的图像增强评价指标

Table 2 Image enhancement evaluation indexes under ${10^{ - 3}}\;{\rm{lx}}$ low-level-light environment

算法	标准差			信息熵			平均梯度
算法	A	B	C	A	B	C	A	B	C
原始图像	21.29	16.71	19.42	5.85	1.81	6.04	0.77	0.69	0.64
交叉子窗口算法	22.39	17.24	19.57	6.04	1.73	6.06	0.87	0.72	0.68
DDE算法	24.67	18.08	19.65	5.87	1.66	5.98	1.06	0.80	0.71
文献[4]算法	26.89	18.70	19.82	6.01	1.63	6.01	0.98	0.87	0.78
本文算法	28.94	19.30	28.43	6.15	1.39	6.39	1.18	0.95	1.13

下载: 导出CSV

表 3 ${10^{ - 5}}\;{\rm{lx}}$微光环境下的图像增强评价指标

Table 3 Image enhancement evaluation indexes under ${10^{ - 5}}\;{\rm{lx}}$ low-level-light environment

算法	标准差			信息熵			平均梯度
算法	A	B	C	A	B	C	A	B	C
原始图像	15.53	9.17	8.08	5.27	1.55	4.55	0.45	0.37	0.26
交叉子窗口算法	14.88	10.10	14.07	5.28	1.38	4.61	0.42	0.43	0.41
DDE算法	17.15	10.57	11.59	5.24	1.30	4.58	0.58	0.44	0.37
文献[4]算法	17.46	11.23	12.02	5.46	1.27	4.61	0.60	0.48	0.42
本文算法	16.57	12.48	15.16	5.21	1.32	3.89	0.63	0.57	0.51

下载: 导出CSV

表 4 4种算法运行时间对比

Table 4 Comparison of running time of four algorithms s

算法	微光环境
算法	$ {10^{ - 1}}\;{\rm{lx}} $	$ {10^{ - 3}}\;{\rm{lx}} $	$ {10^{ - 5}}\;{\rm{lx}} $
交叉子窗口算法	10.371	9.114	6.947
文献[4]算法	9.210	7.062	3.763
本文算法	2.816	2.177	1.041

下载: 导出CSV

参考文献(15)

[1]	徐炜君, 刘国忠. 空间域和频域结合的图像增强技术及实现[J]. 中国测试,2009,35(4):52-54. XU Weijun, LIU Guozhong. Image enhancement technology combining spatial domain and frequency domain and its implementation[J]. China Measurement & Test,2009,35(4):52-54.
[2]	周克虎, 雷涛, 罗刚. 一种基于时域滤波的红外序列图像去噪算法[J]. 应用光学,2021,42(3):474-480. ZHOU Kehu, LEI Tao, LUO Gang. An infrared sequence image denoising algorithm based on time-domain filtering[J]. Journal of Applied Optics,2021,42(3):474-480.
[3]	LIU Ning, ZHAO Dongxue. Detail enhancement for high-dynamic-range infrared images based on guided image filter[J]. Infrared Physics & Technology,2014,67:138-147.
[4]	ZHOU B, LUO Y, YANG M, et al. An improved adaptive detail enhancement algorithm for infrared images based on guided image filter[J]. Journal of Modern Optics,2019,66(1):33-46. doi: 10.1080/09500340.2018.1511861
[5]	王炫, 尹丽菊, 高明亮, 等. 基于新符号函数与盲源分离的光子计数图像去噪方法[J]. 激光与光电子学进展. 2018, 55(10): 158-164. WANG Xuan, YIN Liju, GAO Mingliang, et al. Photon counting image denoising method based on new symbol function and blind source separation[J]. Laser & Optoelectronics Progress. 2018, 55(10): 158-164.
[6]	廖斌, 刘鸳鸳. 基于多尺度灰度变换的图像增强研究[J]. 量子电子学报,2015,32(5):550-554. LIAO Bin, LIU Yuanyuan. Research on image enhancement based on multi-scale gray transform[J]. Chinese Journal of Quantum Electronics,2015,32(5):550-554.
[7]	YIN Hui, GONG Yuanhao, Qiu Guoping. Combined window filtering and its applications[J]. Multidimensional Systems and Signal Processing,2021,32(2):313-333.
[8]	KITAGAWA G. The two-filter formula for smoothing and an implementation of the Gaussian-sum smoother[J]. Annals of the Institute of Statistical Mathematics,1994,46:605-623. doi: 10.1007/BF00773470
[9]	沈德海, 张龙昌, 鄂旭. 基于多子窗口的混合噪声滤波算法[J]. 计算机技术与发展,2015,25(6):69-72. SHEN Dehai, ZHANG Longchang, E Xu. Hybrid noise filtering algorithm based on multiple sub-windows[J]. Computer Technology and Development,2015,25(6):69-72.
[10]	CÁTIA P, ANA O, CRISTINA J, et al. Automatic crackle detection algorithm based on fractal dimension and box filtering[J]. Procedia Computer Science,2015,64:705-712. doi: 10.1016/j.procs.2015.08.592
[11]	胡海波, 邱民仆. 微光图像对比度增强技术研究[J]. 测绘通报,2015,1:26-30. HU Haibo, QIU Minpu. Contrast enhancement of low light level image[J]. Bulletin of Surveying and Mapping,2015,1:26-30.
[12]	孙国栋, 徐昀, 徐亮. 不均匀光照和重影的仪表图像二值化方法[J]. 应用光学,2020,41(1):74-78. doi: 10.5768/JAO202041.0102002 SUN Guodong, XU Yun, XU Liang. Instrumentation image binarization method for inhomogeneous illumination and ghosting[J]. Journal of Applied Optics,2020,41(1):74-78. doi: 10.5768/JAO202041.0102002
[13]	汪伟, 许德海, 任明艺. 一种改进的红外图像自适应增强方法[J]. 红外与激光工程,2021,50(11):419-427. WANG Wei, XU Dehai, REN Mingyi. An improved adaptive enhancement method for infrared image[J]. Infrared and Laser Engineering,2021,50(11):419-427.
[14]	李佳, 李少娟, 段小虎, 等. 基于Retinex理论与概率非局部均值的红外图像增强方法[J]. 光子学报,2020,49(4):187-196. LI jia, LI Shaojuan, DUAN Xiaohu, et al. Infrared image enhancement method based on Retinex theory and probabilistic non-local mean[J]. Acta Photonica Sinica,2020,49(4):187-196.
[15]	汪子君, 罗渊贻, 蒋尚志, 等. 基于引导滤波的自适应红外图像增强改进算法[J]. 光谱学与光谱分析,2020,40(11):3463-3467. WANG Zijun, LUO Yuanyi, JIANG Shangzhi, el at. An improved adaptive infrared image enhancement algorithm based on guided filtering[J]. Spectroscopy and Spectral Analysis,2020,40(11):3463-3467.