基于L1L0层分解模型的家具死节缺陷图像分割
|
摘要
针对成品家具中的表面死节缺陷,提出一种基于L1L0层分解模型与主动轮廓模型的死节缺陷图像分割算法。将RGB彩色图像转换成Lab图像,对Lab各通道灰度图进行L1L0两层分解,得到分解后的低频灰度图与高频系数图;并用主动轮廓模型对各低频灰度图进行分割得到Lab多通道的二值图。最后对多通道的二值图进行比较,选择b通道二值图作为图像最终分割结果,得到死节缺陷目标。试验结果表明,提出的算法能降低复杂背景噪声,分割效果好,能很好地提取家具表面死节缺陷,性能优于传统的Otsu方法。
Aiming at the surface defect of dead knot in finished furniture,an image segmentation algorithm based on L1L0 decomposition model and active contour model was proposed.The RGB color image was transformed into Lab image,and the Lab channel gray images were decomposed by L1L0 two-layer decomposition,hence the decomposed low-frequency gray images and high-frequency coefficients were obtained.Finally,the multi-channel binary image obtained by segmenting these gray images with active contour model was compared.The b-channel binary image was selected as the final segmentation result and the dead-knot defect target was attained.The experimental results show that the proposed algorithm can reduce the noise of complex background,and has good segmentation effect,which can extract the dead knot defects of furniture surface well,and have better performance than the traditional Otsu method.
Aiming at the surface defect of dead knot in finished furniture,an image segmentation algorithm based on L1L0 decomposition model and active contour model was proposed.The RGB color image was transformed into Lab image,and the Lab channel gray images were decomposed by L1L0 two-layer decomposition,hence the decomposed low-frequency gray images and high-frequency coefficients were obtained.Finally,the multi-channel binary image obtained by segmenting these gray images with active contour model was compared.The b-channel binary image was selected as the final segmentation result and the dead-knot defect target was attained.The experimental results show that the proposed algorithm can reduce the noise of complex background,and has good segmentation effect,which can extract the dead knot defects of furniture surface well,and have better performance than the traditional Otsu method.
引文
[1]程玉柱,李赵春.成品家具中的表面死节缺陷图像检测方法[J].家具,2018,39(5):106-110.
[2]祝雅园,吴智慧.数码木纹的扫描采集与色差调整试验研究[J].家具,2018,39(3):8-10,21.
[3]叶雨静,徐伟,黄琼涛,等.实木构件榫接合无损检测技术现状与展望[J].家具,2017,38(5):95-98.
[4]贺雪梅,曹廷蕾.基于形态原型演化的家具产品设计[J].林产工业,2018,45(9):40-43.
[5]王蓓蓓,刘文金,陈耿,等.基于Holt-Winters算法的定制家具生产需求预测研究[J].林产工业,2018,45(1):44-47.
[6]郭颖艳,杨凌云.数字化技术在川西传统木作传承创新中的应用[J].林产工业,2018,45(11):70-73.
[7]程玉柱,段一凡,李赵春.基于PSO与水平集的桂花图像分割方法[J].中国农机化学报,2018,39(8):81-84,91.
[8]袁雯雯,姜树海.林火图像识别理论研究进展[J].世界林业研究,2018,31(1):35-39.
[9]程玉柱,陈勇,张浩.基于MMC与CV模型的苗期玉米图像分割算法[J].农业机械学报,2013,44(11):266-270.
[10]周咏,顾桂梅,冉建民.基于改进主动轮廓模型的钢轨表面缺陷分割[J].兰州交通大学学报,2018,37(3):49-57.
[11]张佳明,张荣国,赵建,等.基于纹理特征的主动轮廓模型图像分割[J].太原科技大学学报,2017,38(3):190-195.
[12]程玉柱,李赵春.基于改进的Bingham模型的图像去噪算法[J].电子技术应用,2018,44(6):72-75.
[13] HAMMOND D L. Validation of LAB color mode as a non-destructive method to differentiate black ballpoint pen inks.[J]. Journal of Forensic Sciences,2007,52(4):7.
[14] LIANG Z T,XU J,ZHANG D,et al. A Hybrid L1-L0Layer Decomposition Model for Tone Mapping[C]. Computer Vision and Pattern Recognition(CVPR),Salt Lake City,USA,2018.
[15]刘传泽,罗瑞,陈龙现,等.基于区域筛选分割和随机森林的人造板表面缺陷识别[J].制造业自动化,2018,40(9):9-13.
[16]王程程,刘亚秋.基于动态目标检测与跟踪算法(TLD)的刨花板表面缺陷检测技术[J].木材工业,2018,32(4):44-47.
[2]祝雅园,吴智慧.数码木纹的扫描采集与色差调整试验研究[J].家具,2018,39(3):8-10,21.
[3]叶雨静,徐伟,黄琼涛,等.实木构件榫接合无损检测技术现状与展望[J].家具,2017,38(5):95-98.
[4]贺雪梅,曹廷蕾.基于形态原型演化的家具产品设计[J].林产工业,2018,45(9):40-43.
[5]王蓓蓓,刘文金,陈耿,等.基于Holt-Winters算法的定制家具生产需求预测研究[J].林产工业,2018,45(1):44-47.
[6]郭颖艳,杨凌云.数字化技术在川西传统木作传承创新中的应用[J].林产工业,2018,45(11):70-73.
[7]程玉柱,段一凡,李赵春.基于PSO与水平集的桂花图像分割方法[J].中国农机化学报,2018,39(8):81-84,91.
[8]袁雯雯,姜树海.林火图像识别理论研究进展[J].世界林业研究,2018,31(1):35-39.
[9]程玉柱,陈勇,张浩.基于MMC与CV模型的苗期玉米图像分割算法[J].农业机械学报,2013,44(11):266-270.
[10]周咏,顾桂梅,冉建民.基于改进主动轮廓模型的钢轨表面缺陷分割[J].兰州交通大学学报,2018,37(3):49-57.
[11]张佳明,张荣国,赵建,等.基于纹理特征的主动轮廓模型图像分割[J].太原科技大学学报,2017,38(3):190-195.
[12]程玉柱,李赵春.基于改进的Bingham模型的图像去噪算法[J].电子技术应用,2018,44(6):72-75.
[13] HAMMOND D L. Validation of LAB color mode as a non-destructive method to differentiate black ballpoint pen inks.[J]. Journal of Forensic Sciences,2007,52(4):7.
[14] LIANG Z T,XU J,ZHANG D,et al. A Hybrid L1-L0Layer Decomposition Model for Tone Mapping[C]. Computer Vision and Pattern Recognition(CVPR),Salt Lake City,USA,2018.
[15]刘传泽,罗瑞,陈龙现,等.基于区域筛选分割和随机森林的人造板表面缺陷识别[J].制造业自动化,2018,40(9):9-13.
[16]王程程,刘亚秋.基于动态目标检测与跟踪算法(TLD)的刨花板表面缺陷检测技术[J].木材工业,2018,32(4):44-47.