烧结机火焰图像的小波去噪及亮度特征分析
|
摘要
烧结机尾矿料断面火焰图像中蕴含着大量的与烧结终点关联的信息,但图像中含有的噪声及光晕干扰影响了对图像信息的利用,为此采用多尺度小波阈值方法分别对火焰断面图像和图像亮度变化曲线进行去噪处理。首先,连续采集机尾烧结矿料断面火焰视频并转化为1 522帧序列图像;其次,采用多尺度小波阈值分解方法提取火焰目标图像的小波系数,对图像进行去噪、分割处理;再次,提取滤波后的火焰图像亮度特征,得到火焰序列图像的连续亮度变化曲线,并通过计算不同区间序列亮度曲线下的积分面积对火焰燃烧状态进行分析;最后,对火焰图像亮度曲线再次进行小波分解、系数提取及信号重构,获得平滑的火焰亮度特征变化曲线。
There is a large amount of information related to burning through point(BTP)in the flame image of sinter section,however,the noise and halo interference in the image affects the use of the image information,so multi-scale wavelet threshold method was used to de-noise the flame section image and brightness curve of flame section image.Firstly,the flame video of the sinter section was collected and transformed into 1 522 frame image.Secondly,the wavelet coefficients of flame target image were extracted by multi-scale wavelet threshold decomposition method,and the image was processed with de-noising and segmentation.Thirdly,the brightness characteristics of the flame image were extracted,and the continuous brightness curve of the flame sequence image was obtained,and the combustion state of flame was analyzed by calculating the integral area under brightness curves.Finally,the flame image brightness curve was decomposed by wavelet transform,coefficient extraction and signal reconstruction,and the characteristic curve of smooth flame brightness was obtained.
There is a large amount of information related to burning through point(BTP)in the flame image of sinter section,however,the noise and halo interference in the image affects the use of the image information,so multi-scale wavelet threshold method was used to de-noise the flame section image and brightness curve of flame section image.Firstly,the flame video of the sinter section was collected and transformed into 1 522 frame image.Secondly,the wavelet coefficients of flame target image were extracted by multi-scale wavelet threshold decomposition method,and the image was processed with de-noising and segmentation.Thirdly,the brightness characteristics of the flame image were extracted,and the continuous brightness curve of the flame sequence image was obtained,and the combustion state of flame was analyzed by calculating the integral area under brightness curves.Finally,the flame image brightness curve was decomposed by wavelet transform,coefficient extraction and signal reconstruction,and the characteristic curve of smooth flame brightness was obtained.
引文
[1]龙红明.铁矿粉烧结原理与工艺[M].北京:冶金工业出版社,2010.
[2]王培珍,郑诗程,周芳,等.基于模糊推理与神经网络的烧结图像识别[J].烧结球团,2006,31(6):22-27.
[3]吴晶.基于烧结机尾断面图像的烧结质量人工智能检测方法的研究[D].沈阳:东北大学,2013:28-40.
[4] Tahara T,Shimobaba T,Ito T.Image-reconstruction algorithm with no use of Fourier transform in interferometric imaging using spatial frequency-division multiplexing[C]//Adaptive Optics:Analysis,Methods&Systems.USA:Optical Society of America,2016.
[5] Zhang Y D,Chen S,Wang S H,et al.Magnetic resonance brain image classification based on weighted-type fractional Fourier transform and nonparallel support vector machine[J].International Journal of Imaging Systems and Technology,2015,25(4):317-327.
[6]张霞,黄继风.结合LBP直方图和SVM的视频火焰检测[J].计算机应用与软件,2016,33(8):216-220.
[7] Lima J B,Novaes L F G.Image encryption based on the fractional Fourier transform over finite fields[J].Signal Processing,2014,94:521-530.
[8] Barron J T,Malik J.Intrinsic scene properties from a single RGB-D image[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.[s.l.]:IEEE,2013:17-24.
[9] Chakrabarti A,Sunkavalli K.Single-image RGB photometric stereo with spatially-varying albedo[C]//2016 Fourth International Conference on 3DVision.[s.l.]:IEEE,2016.
[10] Wang S,Pan H,Zhang C,et al.RGB-D image-based detection of stairs,pedestrian crosswalks and traffic signs[J].Journal of Visual Communication and Image Representation,2014,25(2):263-272.
[2]王培珍,郑诗程,周芳,等.基于模糊推理与神经网络的烧结图像识别[J].烧结球团,2006,31(6):22-27.
[3]吴晶.基于烧结机尾断面图像的烧结质量人工智能检测方法的研究[D].沈阳:东北大学,2013:28-40.
[4] Tahara T,Shimobaba T,Ito T.Image-reconstruction algorithm with no use of Fourier transform in interferometric imaging using spatial frequency-division multiplexing[C]//Adaptive Optics:Analysis,Methods&Systems.USA:Optical Society of America,2016.
[5] Zhang Y D,Chen S,Wang S H,et al.Magnetic resonance brain image classification based on weighted-type fractional Fourier transform and nonparallel support vector machine[J].International Journal of Imaging Systems and Technology,2015,25(4):317-327.
[6]张霞,黄继风.结合LBP直方图和SVM的视频火焰检测[J].计算机应用与软件,2016,33(8):216-220.
[7] Lima J B,Novaes L F G.Image encryption based on the fractional Fourier transform over finite fields[J].Signal Processing,2014,94:521-530.
[8] Barron J T,Malik J.Intrinsic scene properties from a single RGB-D image[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.[s.l.]:IEEE,2013:17-24.
[9] Chakrabarti A,Sunkavalli K.Single-image RGB photometric stereo with spatially-varying albedo[C]//2016 Fourth International Conference on 3DVision.[s.l.]:IEEE,2016.
[10] Wang S,Pan H,Zhang C,et al.RGB-D image-based detection of stairs,pedestrian crosswalks and traffic signs[J].Journal of Visual Communication and Image Representation,2014,25(2):263-272.