基于通道差值模型的导向滤波去雾算法及其FPGA实现

doi:10.16180/j.cnki.issn1007-7820.2023.08.001

电子科技 ›› 2023, Vol. 36 ›› Issue (8): 1-6.doi: 10.16180/j.cnki.issn1007-7820.2023.08.001

• • 下一篇

基于通道差值模型的导向滤波去雾算法及其FPGA实现

曹红芳,王晓蕾,杜高明,李桢旻,倪伟

合肥工业大学微电子设计研究所,安徽合肥 230601

收稿日期:2022-03-17 出版日期:2023-08-15 发布日期:2023-08-14
作者简介:曹红芳(1995-),女,硕士研究生。研究方向:图像处理和集成电路设计。|王晓蕾(1978-),女,博士,副教授。研究方向:集成电路设计理论。|杜高明(1977-),男,博士,教授。研究方向:多核体系结构、片上网络体系结构。|李桢旻(1982-),男,博士,讲师。研究方向:时间可预测软硬件设计、众核片上网络技术。|倪伟(1977-),男,博士,副教授。研究方向:数字集成电路设计、可重构计算、人工智能、深度学习。
基金资助:
国家重点研发计划(2018YFB2202604);安徽省高校协同创新项目(GXXT-2019-030)

Design and FPGA Implementation of Dehazing Based on Channel Difference Model and Guided Filtering

CAO Hongfang,WANG Xiaolei,DU Gaoming,LI Zhenmin,NI Wei

Institute of VLSI Design,Hefei University of Technology,Hefei 230601,China

Received:2022-03-17 Online:2023-08-15 Published:2023-08-14
Supported by:
National Key R&D Program of China(2018YFB2202604);University Synergy Innovation Program of Anhui(GXXT-2019-030)

摘要/Abstract

摘要：

计算机视觉系统受到有雾天气的影响会导致捕获的图像质量较差。为了解决该问题,文中提出了一种基于通道差值模型的导向滤波去雾算法及其FPGA(Field Programmable Gate Array)设计。通过分离雾天图像的亮通道与暗通道得到通道差值模型,并将该模型作为导向滤波的引导图对雾天图像进行平滑处理,最后进行高升压滤波操作得到去雾图像,设计硬件架构并在FPGA上实现。实验结果表明,去雾后的图像场景照度均匀,纹理信息恢复程度较好且颜色保真度高,对于480×270大小的图像,电路综合频率为108.448 MHz,吞吐量为323.47 MB·s^-1,完成整个去雾过程花费时间为0.001 2 s。实验结果证明文中所提算法及其硬件设计能够有效提高图像可见度和去雾速度。

关键词: 去雾, 通道差值模型, 导向滤波, 高升压滤波, 频率, 吞吐量, 图像处理, FPGA

Abstract:

Computer vision systems are affected by foggy weather, resulting in poor quality images captured. To solve this problem, this study proposes a guided filtering dehazing algorithm based on channel difference model and its FPGA design. The channel difference model is obtained by separating the bright channel and dark channel of foggy image, and the model is used as a guide map for guided filtering to smooth the foggy image. Finally, a high boost filtering operation is performed to obtain a dehazed image. The hardware architecture is designed and implemented on FPGA. The experimental results show that the image scene after dehazing has uniform illumination, high degree of texture information recovery and high color fidelity. For an image of 480×270 size, the integrated frequency of the circuit is 108.448 MHz, the throughput is 323.47 MB·s^-1, and the time to complete the entire dehazing is 0.001 2 s. These results indicate that the proposed algorithm and its hardware design can effectively improve image visibility and dehazing speed.

Key words: dehazing, channel difference model, guided filtering, high boost filter, frequency, throughput, image processing, FPGA

中图分类号:

TP391.41

曹红芳,王晓蕾,杜高明,李桢旻,倪伟. 基于通道差值模型的导向滤波去雾算法及其FPGA实现[J]. 电子科技, 2023, 36(8): 1-6.

CAO Hongfang,WANG Xiaolei,DU Gaoming,LI Zhenmin,NI Wei. Design and FPGA Implementation of Dehazing Based on Channel Difference Model and Guided Filtering[J]. Electronic Science and Technology, 2023, 36(8): 1-6.

图/表 10

图1

图2

图3

图4

图5

图6

表1

不同场景下各去雾方法的定量比较"

图像	参数	文献[14]	文献[15]	本文
图6(a)	e	5.582 6	3.936 7	6.458 3
	$r -$	3.889 2	1.765 2	4.940 0
	σ	0.000 0	0.000 0	0.000 3
图6(b)	e	6.468 4	7.055 6	6.706 8
	$r -$	2.767 8	1.888 3	4.420 8
	σ	0.079 6	0.000 5	0.032 4
图6(c)	e	8.523 7	4.460 1	14.072 1
	$r -$	2.351 0	2.371 8	3.580 6
	σ	0.033 2	0.000 2	0.098 6
图6(d)	e	15.088 8	14.230 8	17.732 4
	$r -$	2.537 5	2.100 8	4.1210
	σ	0.003 7	0.000 0	0.020 2

表1

表2

表3

表4

参考文献 18

[1]	董静薇, 赵春丽, 海博. 融合同态滤波和小波变换的图像去雾算法研究[J]. 哈尔滨理工大学学报, 2019, 24(1):66-77.
	Dong Jingwei, Zhao Chunli, Hai Bo. Research on image de-fog algorithm based on fusion homomorphic filtering and wavelet transform[J]. Journal of Harbin University of Science and Technology, 2019, 24(1):66-77.
[2]	董丽丽, 丁畅, 许文海. 基于直方图均衡化图像增强的两种改进方法[J]. 电子学报, 2018, 46(10):2367-2375. doi: 10.3969/j.issn.0372-2112.2018.10.009
	Dong Lili, Ding Chang, Xu Wenhai. Two improved methods based on histogram equalization for image enhancement[J]. Acta Electronica Sinca, 2018, 46(10):2367-2375. doi: 10.3969/j.issn.0372-2112.2018.10.009
[3]	Xu X G, Yang P, Liu Y L. Night image dehazing based on full-scale retinex algorithm[J]. Microelectronics & Computer, 2017, 34(7):132-136.
[4]	Cai B, Xu X, Jia K, et al. DehazeNet: An end-to-end system for single image haze removal[J]. IEEE Transactions on Image Processing, 2016, 25(11):5187-5198. doi: 10.1109/TIP.2016.2598681 pmid: 28873058
[5]	何宜鸿, 李彦锋, 黄树恺, 等. 基于深度卷积神经网络的自适应图像去雾算法[J]. 电子科技, 2020, 33(8):70-73.
	He Yihong, Li Yanfeng, Huang Shukai, et al. Adaptive image dehazing algorithm based on deep convolutional neural network[J]. Electronic Science and Technology, 2020, 33(8):70-73.
[6]	Kim Y, Yim C. Image dehaze method using depth map estimation network based on atmospheric scattering model[C]. Barcelona: Proceedings of the International Conference on Electronics, Information, and Communication, 2020:1-3.
[7]	Choi H, Lim H, Yu S, et al. Haze removal of multispectral remote sensing imagery using atmospheric scattering model-based haze thickness map[C]. Seoul: Proceedings of the IEEE International Conference on Consumer Electronics-Asia, 2020:1-2.
[8]	He K, Sun J, Tang X. Single image haze removal using dark channel prior[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010, 33(12):2341-2353. doi: 10.1109/TPAMI.2010.168
[9]	Vazquez-Corral J, Finlayson G D, Bertalmío M. Physical-based optimization for non-physical image dehazing methods[J]. Optics Express, 2020, 28(7):9327-9339. doi: 10.1364/OE.383799 pmid: 32225542
[10]	李喆, 李建增, 胡永江, 等. 混合先验与加权引导滤波的图像去雾算法[J]. 中国图象图形学报, 2019, 24(2):170-179.
	Li Zhe, Li Jianzeng, Hu Yongjiang, et al. Image dehazing algorithm based on hybrid prior and weighted guided filtering[J]. Journal of Image and Graphics, 2019, 24(2):170-179.
[11]	陆海俊, 汪荣贵, 杨娟, 等. 基于均值漂移的暗原色先验图像去雾算法[J]. 合肥工业大学学报(自然科学版), 2016, 39(9):1205-1210.
	Lu Haijun, Wang Ronggui, Yang Juan, et al. An image haze removal algorithm using dark prior based on mean shift[J]. Journal of Harbin University of Science and Technology, 2016, 39(9):170-179.
[12]	Shiau Y H, Kuo Y T, Chen P Y, et al. VLSI design of an efficient flicker-free video defogging method for real-time applications[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2017, 29(1):238-251. doi: 10.1109/TCSVT.2017.2777140
[13]	Singh R, Biswas M. Contrast and color improvement based haze removal of underwater images using fusion technique[C]. Solan: Proceedings of the Fourth International Conference on Signal Processing, Computing and Control, 2017:138-143.
[14]	Du G, Wu J, Cao H, et al. A real-time effective fusion-based image defogging architecture on FPGA[J]. ACM Transactions on Multimedia Computing, Communications, and Applications, 2021, 17(3):1-21.
[15]	Galdran A. Image dehazing by artificial multiple-exposure image fusion[J]. Signal Processing, 2018, 14(9):135-147.
[16]	Koga T, Yasuda A, Furukawa S, et al. A simplified and fast dehazing processing suitable for embedded systems[C]. Ishigaki: Proceedings of the International Symposium on Intelligent Signal Processing and Communication Systems, 2018:492-497.
[17]	Khurshid L, Raja G. Real time architecture for image dehazing[C]. Online: Proceedings of the Eleventh International Conference of Pattern Recognition Systems, 2021:79-84.
[18]	Zhao Z Y, Xiao A T, Guo J I. Video dehazing hardware accelerator design based on dark channel prior with sky preservation[C]. Hsinchu: Proceedings of the International Symposiumon VLSI Design, Automation and Test, 2020:1-4.

FPGA逻辑器件	设计利用总数
FPGA逻辑器件	已用	可用	占比/%
LUT总数	44 690	203 800	21.93
LUTRAM总数	33 288	64 000	52.01
触发器总数	11 450	407 600	2.81
DSP块总数	69	840	8.21
频率/MHz	-	108.448	-
时钟周期总数	-	130 351	-
吞吐量/MB·s^-1	-	323.47	-
时间/s	-	0.001 2	-

实现平台	计算时间/s	加速倍率
软件	0.440 0	1
硬件	0.001 2	367

图像大小	文献[16]/s	文献[17]/s	文献[18]/s	本文/s
640×480	0.025	0.064	0.013	0.003
1 024×768	0.058	0.150	0.031	0.007
1 280×960	0.092	0.235	0.048	0.011
1 920×1 080	0.158	0.407	0.084	0.019

基于通道差值模型的导向滤波去雾算法及其FPGA实现

Design and FPGA Implementation of Dehazing Based on Channel Difference Model and Guided Filtering

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 18

相关文章 15

Metrics

本文评价

推荐阅读 0

[1]	顾鑫,许伯强,徐桂东,徐晨光,张赛. 基于Bloch-Floquent理论和有限元仿真的板结构Lamb波复频散关系研究[J]. 电子科技, 2023, 36(8): 49-55.
[2]	查俊伟,张洪艳. 动态感受野特征选择去雾网络[J]. 电子科技, 2023, 36(7): 56-63.
[3]	郭齐成,沈拓,张轩雄. 基于图像处理的轨旁组合信号灯识别方法[J]. 电子科技, 2023, 36(7): 8-15.
[4]	苑振博,白勃,张晓薇,罗柳君,尚韬. 基于图像处理的可见光通信自动对准系统[J]. 电子科技, 2023, 36(6): 8-15.
[5]	朱家文,陈卓,刘人志,刘柏霖,陈湘萍. 风电机组惯量支撑与一次调频综合控制策略[J]. 电子科技, 2023, 36(5): 9-15.
[6]	罗欣,冯武,孙卫杰,刘马良. 基于FPGA的多功能阵列信号处理系统设计[J]. 电子科技, 2023, 36(3): 1-6.
[7]	赵英潇,苏阳. 基于FPGA的多通道雷达数据回放系统设计[J]. 电子科技, 2023, 36(2): 1-7.
[8]	刘美红,洪恩航,黎振华,滕宝仁. 基于图像处理与数值仿真辅助的电弧增材成形研究[J]. 电子科技, 2023, 36(1): 7-14.
[9]	徐邦贤,刘晓波,韩祥民,邱知,唐辉,范津玮. 耦合电感型Zeta变换器的参数优化方法[J]. 电子科技, 2023, 36(1): 88-94.
[10]	毕嘉桢,沈拓,张轩雄. 基于机器视觉的轨道交通自动测距研究[J]. 电子科技, 2022, 35(9): 37-43.
[11]	冯俊逸,沈拓,张轩雄. 基于图像处理的轨旁信号机识别方法[J]. 电子科技, 2022, 35(8): 53-57.
[12]	孙翔,裴晓芳,周望,朱平. 基于标尺图像识别的作物株高测量[J]. 电子科技, 2022, 35(7): 32-39.
[13]	张崇崇,黄亚宇. GA-BP神经网络对片烟结构的预测研究[J]. 电子科技, 2022, 35(6): 35-42.
[14]	杨锐,孙岩洲,余江华,张蒙飞,韦延方. 电场耦合式无线电能传输频率分裂特性分析[J]. 电子科技, 2022, 35(6): 48-53.
[15]	王新,黄冲,许翔. PC三电平逆变器电解电容故障特征分析[J]. 电子科技, 2022, 35(5): 81-86.