基于裂变粒子滤波算法的织物图像疵点检测研究
|
摘要
为了提高织物图像疵点检测的效果采用裂变粒子滤波算法。通过分类复制算法对粒子进行选择,整个过程粒子总数不变;裂变因子控制粒子裂变数量与其对应的被裂变粒子权值成正比,大权值的粒子能够裂变生成更多的粒子,根据多样性函数以及广义似然比检验定律判断是否处于有限收敛界,若是则停止裂变;织物图像的像素点分预测、更新消噪过程,疵点区域通过最佳阈值分割;给出织物图像疵点检测过程。实验仿真显示,此算法对织物图像疵点检测效果清晰,疵点在整体上保持了较为完整的检测效果,误检率、检出率指标较优。
In order to improve the effect of fabric image defect detection, fission particle filtering algo?rithm was proposed. Firstly, particle was selected by classification and replication algorithm, the total number of particles in the whole process was unchanged. Secondly, the number of fission particles was controlled by fission factor to be proportional to the weight of the corresponding fission particles, particles in the large weight value would fissure to produce more particles. The diversity function and the generalized likelihood ra?tio law were used to judge whether it was in the finite convergence bound. If it was, stop fission. Thirdly, the image denoising process of the fabric was divided into forecast and update, and the defect area was segment?ed through the best threshold. Finally, the detection process of the image defect point of the fabric was given.Experimental simulation showed that the algorithm had a clearer detection effect on the fabric image defects,the defect had complete detection effect on the whole, and the error detection rate and detection rate were better than other algorithms.
In order to improve the effect of fabric image defect detection, fission particle filtering algo?rithm was proposed. Firstly, particle was selected by classification and replication algorithm, the total number of particles in the whole process was unchanged. Secondly, the number of fission particles was controlled by fission factor to be proportional to the weight of the corresponding fission particles, particles in the large weight value would fissure to produce more particles. The diversity function and the generalized likelihood ra?tio law were used to judge whether it was in the finite convergence bound. If it was, stop fission. Thirdly, the image denoising process of the fabric was divided into forecast and update, and the defect area was segment?ed through the best threshold. Finally, the detection process of the image defect point of the fabric was given.Experimental simulation showed that the algorithm had a clearer detection effect on the fabric image defects,the defect had complete detection effect on the whole, and the error detection rate and detection rate were better than other algorithms.
引文
[1]樊翠玲.基于改进粒子群算法的织物图像疵点分割研究[J].毛纺科技,2018,46(4):55-60.
[2]朱金华,李玉仙.基于卡尔曼滤波的图像降噪算法[J].现代电子技术,2017,40(9):22-24.
[3]游斌相,刘桂华,王姮.基于小波变换的磁粉缺陷图像特征提取与检测[J].自动化仪表,2017,38(12):89-93.
[4]魏彦.基于自适应均值的图像滤波算法[J].包装工程,2017,38(13):204-207.
[5]杨超宇,李策,梁胤程,等.基于改进粒子滤波的煤矿视频监控模糊目标检测[J].吉林大学学报:工学版,2017,47(6):1 976-1 985.
[6]刘梦菱,秦岭.基于正则化的高斯粒子滤波算法[J].计算技术与自动化,2014,33(1):69-72.
[7]杨伟博,袁慎芳,邱雷,等.基于辅助粒子滤波的疲劳裂纹扩展预测研究[J].振动与冲击,2018,37(5):114-119.
[8]高国栋,林明,许兰.基于似然分布调整的粒子群优化粒子滤波新方法[J].计算机应用,2017,37(4):980-985.
[9]朱福利,曾碧,曹军.基于粒子滤波的SLAM算法并行优化与实现[J].广东工业大学学报,2017,34(2):92-96.
[10]危璋,冯新喜,毛少锋.裂变优选重采样粒子滤波算法[J].空军工程大学学报:自然科学版,2014,15(6):82-86.
[11]王泽玉,李明,张鹏.一种新的进化裂变粒子滤波算法[J].西安电子科技大学学报,2014,12(6):31-36.
[12]韩雪.自适应裂变粒子滤波压制地震勘探随机噪声[D].长春:吉林大学,2016.
[13]赛地瓦尔地·买买提,阿布都加怕尔·如苏力,米娜瓦尔·吾买尔.基于卡尔曼滤波的图像降噪方法研究[J].通信技术,2016,49(4):423-425.
[14]刘海军,单维锋,袁静,等.基于梯度方向直方图的本色布疵点检测算法[J].毛纺科技,2018,46(1):69-72.
[2]朱金华,李玉仙.基于卡尔曼滤波的图像降噪算法[J].现代电子技术,2017,40(9):22-24.
[3]游斌相,刘桂华,王姮.基于小波变换的磁粉缺陷图像特征提取与检测[J].自动化仪表,2017,38(12):89-93.
[4]魏彦.基于自适应均值的图像滤波算法[J].包装工程,2017,38(13):204-207.
[5]杨超宇,李策,梁胤程,等.基于改进粒子滤波的煤矿视频监控模糊目标检测[J].吉林大学学报:工学版,2017,47(6):1 976-1 985.
[6]刘梦菱,秦岭.基于正则化的高斯粒子滤波算法[J].计算技术与自动化,2014,33(1):69-72.
[7]杨伟博,袁慎芳,邱雷,等.基于辅助粒子滤波的疲劳裂纹扩展预测研究[J].振动与冲击,2018,37(5):114-119.
[8]高国栋,林明,许兰.基于似然分布调整的粒子群优化粒子滤波新方法[J].计算机应用,2017,37(4):980-985.
[9]朱福利,曾碧,曹军.基于粒子滤波的SLAM算法并行优化与实现[J].广东工业大学学报,2017,34(2):92-96.
[10]危璋,冯新喜,毛少锋.裂变优选重采样粒子滤波算法[J].空军工程大学学报:自然科学版,2014,15(6):82-86.
[11]王泽玉,李明,张鹏.一种新的进化裂变粒子滤波算法[J].西安电子科技大学学报,2014,12(6):31-36.
[12]韩雪.自适应裂变粒子滤波压制地震勘探随机噪声[D].长春:吉林大学,2016.
[13]赛地瓦尔地·买买提,阿布都加怕尔·如苏力,米娜瓦尔·吾买尔.基于卡尔曼滤波的图像降噪方法研究[J].通信技术,2016,49(4):423-425.
[14]刘海军,单维锋,袁静,等.基于梯度方向直方图的本色布疵点检测算法[J].毛纺科技,2018,46(1):69-72.