BM3D去噪算法在仪表图像识别中的应用Application of BM3D Algorithm in Instrument Image Recognition
陈波,丁宁,边境,孙慧媛
CHEN Bo,DING Ning,BIAN Jing,SUN Huiyuan
摘要(Abstract)：
电力系统中巡检机器人在采集仪表图像时往往受噪声影响而将有用信息掩盖。为解决这一问题,将BM3D(三维块匹配)算法应用于仪表图像的处理中,能够得到较好的去噪效果。BM3D算法是基于变换阈方法和非局部思想发展而来的。首先通过块匹配生成三维矩阵,然后在三维变化域去噪,最后逆变换还原图像。通过实验对BM3D去噪方法与其他传统方法的性能进行了对比分析。实验结果表明, BM3D图像去噪算法可以较好地保留边缘细节效果,解决了恢复图像细节与抑制噪声产生之间的矛盾,且效率较高。
When the inspection robot is applied in the power system, the noise often affects the collected image of the instrument and conceals useful information. In order to solve this problem, this paper applies BM3D(block-matching and 3D filtering) algorithm to the filtering of the instrument image and gets a better denoising effect. The BM3D algorithm derives from transform threshold methods and non-local ideas. The method first generates a three-dimensional matrix through block matching, then denoises in the three-dimensional change domain and finally inversely transforms the restored image. This paper analyzes the BM3D image denoising algorithm and compares the performance of BM3 D denoising method with other traditional image denoising methods through the denoising experiment of instrument images. The experimental results show that the BM3D image denoising algorithm can maintain a good edge detail effect and solves the conflict between the restoration of image details and the suppression of noise, and has higher efficiency.
关键词(KeyWords)：
BM3D;图像去噪;块匹配;三维变换
BM3D;image denoising;block matching;3D transform
基金项目(Foundation):
作者(Author):
陈波,丁宁,边境,孙慧媛
CHEN Bo,DING Ning,BIAN Jing,SUN Huiyuan
DOI: 10.19585/j.zjdl.201903010
参考文献(References)：
- [1]DABOV K,FOI A,KATKOVNIK V,et al.Image denoising by sparse 3-D transform-domain collaborative filtering[J].IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society,2007,16(8):2080.
- [2]李政,刘文江,戎蒙恬,等.BM3D视频去噪算法实现与评估[J].信息技术,2012(4):30-32.
- [3]黄牧,黄文清,李俊柏,等.基于BM3D图像去噪算法的参数研究[J].工业控制计算机,2014(10):99-101.
- [4]MAGGIONI M,FOI A.Joint removal of random and fixedpattern noise through spatiotemporal video filtering[J].IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society,2014,23(10):4282-4296.
- [5]高陈强,李佩.引导滤波和三维块匹配结合的红外图像去噪[J].重庆邮电大学学报(自然科学版),2016,28(2):150-155.
- [6]石健,汪洋,黄海风,等.BM3D算法在海洋SAR图像去噪中的应用[J].雷达科学与技术,2016,14(1):24-32.
- [7]刘思延.基于自适应阈值的三维块匹配降噪算法研究[D].广州:中北大学,2016.
- [8]杨东盛.基于剪切波变换和图像块匹配的图像融合算法研究[D].北京:北京交通大学,2017.
- [9]杨成佳.图像去噪及其效果评估若干问题研究[D].吉林大学,2016.
- [10]赵晓雷.基于小波变换和均值滤波的图像去噪研究[J].信息技术,2017(2):69-71.
- [11]陈舫明,杜雅慧,陈弘.图像处理技术在输电线路直升机智能巡检中的应用[J].浙江电力,2012,31(9):63-66.
- [12]邹成伍,吴剑芳,吕几凡,等.基于图像识别的数字多用表自动化检定系统关键技术及实现[J].浙江电力,2017,36(4):14-17.
- [13]欧家祥,史文彬,张俊玮,等.基于深度学习的高效电力部件识别[J].电力大数据,2018,21(9):1-8.
- [14]郭圣,曾懿辉,张纪宾,等.输电线路防外力破坏智能监控系统的应用[J].广东电力,2018,31(4):139-143.
- [15]康乐,冯殿义,胡2)华.液体泡沫中单气泡的图像识别算法[J].渤海大学学报(自然科学版),2017,38(2):167-172.