多阈值优化的运动图像轮廓特征提取方法
|
摘要
针对传统运动图像轮廓特征提取方法存在提取时间较长、提取精度较低的问题,提出多阈值优化的运动图像轮廓特征提取方法.通过运动图像轮廓特征分析,利用最大类间方差模糊约束法获取隶属度函数,利用模糊隶属计算运动图像中目标体轮廓多个阈值,利用约束后的多个阈值计算图像轮廓范围内中心点近邻的两个轮廓点的几何中心值,通过计算曲率角得到曲率符号,依据曲率符号提取运动图像轮廓特征.结果表明,所提方法曲率特征计算误差率低,有效减少了运动图像轮廓特征提取时间,提高了特征信息提取精度.
In order to solve the problem of long extraction time and lowextraction precision related to the traditional extraction methods of contour feature for motion images,an extraction method of contour features by multi-threshold optimization for motion images was proposed. Through the analysis on the contour features for motion images,the membership function was obtained with the maximum inter-class variance fuzzy constraint method. Multiple thresholds of object contour for motion images were calculated with the fuzzy membership. The geometric center values of two contour points near the center point within the image contour were calculated with the multiple constrained thresholds. The curvature symbol was obtained through calculating curvature angles,and the contour features of motion images were extracted according to the curvature symbol. The results showthat the proposed method has a lower calculation error rates of curvature feature,reducing the extraction time of contour features for motion images effectively and enhancing the extraction accuracy of feature information.
In order to solve the problem of long extraction time and lowextraction precision related to the traditional extraction methods of contour feature for motion images,an extraction method of contour features by multi-threshold optimization for motion images was proposed. Through the analysis on the contour features for motion images,the membership function was obtained with the maximum inter-class variance fuzzy constraint method. Multiple thresholds of object contour for motion images were calculated with the fuzzy membership. The geometric center values of two contour points near the center point within the image contour were calculated with the multiple constrained thresholds. The curvature symbol was obtained through calculating curvature angles,and the contour features of motion images were extracted according to the curvature symbol. The results showthat the proposed method has a lower calculation error rates of curvature feature,reducing the extraction time of contour features for motion images effectively and enhancing the extraction accuracy of feature information.
引文
[1]张婷曼.基于小波阈值自适应修正的模糊图像修复算法[J].控制工程,2015,22(6):1166-1170.(ZHANG Ting-man.Fuzzy image restoration algorithm based on w avelet threshold adaptive correction[J].Control Engineering of China,2015,22(6):1166-1170.)
[2]王丽梅,张宗雪.H型精密运动平台交叉耦合模糊PID同步控制[J].沈阳工业大学学报,2018,40(1):1-5.(WANG Li-mei,ZHANG Zong-xue.Cross-coupling fuzzy PID synchronization control for H precision motion platform[J].Journal of Shenyang University of Technology,2018,40(1):1-5.)
[3]刘小凯,秦川.基于图像块分类阈值优化的改进可逆图像伪装[J].应用科学学报,2018,36(2):237-246.(LIU Xiao-kai,QIN Chuan.Improved reversible image camouflage based on threshold optimization of image block classification[J].Journal of Applied Science,2018,36(2):237-246.)
[4]王文豪,严云洋,姜明新,等.一种去噪声的轮廓提取算法[J].江苏科技大学学报(自然科学版),2017,31(4):519-524.(WANG Wen-hao,YAN Yun-yang,JIANG Mingxin,et al.A noise removing contour extraction algorithm[J].Journal of Jiangsu University of Science and Technology(Natural Science Edition),2017,31(4):519-524.)
[5]韩立博,马安然,任宁宁,等.基于轮廓和灰度特征匹配的X线影像成像部位自动判定方法研究[J].中国医疗器械杂志,2016,40(5):318-321.(HAN Li-bo,MA An-ran,REN Ning-ning,et al.The automatic X-ray radiological positions detection based on contour and gary feature matching[J].Chinese Journal of M edical Lnstrumentation,2016,40(5):318-321.)
[6]张恒,徐刚,刘艳丽,等.基于结构森林的RGB-D图像轮廓提取[J].计算机应用研究,2017,34(2):631-635.(ZHANG Heng,XU Gang,LIU Yan-li,et al.Contour extraction from RGB-D image based on structured forests[J].Application Research of Computers,2017,34(2):631-635.)
[7]王昊,彭博,陈琴,等.基于多尺度融合的甲状腺结节图像特征提取[J].数据采集与处理,2016,31(5):1004-1009.(WANG Hao,PENG Bo,CHEN Qin,et al.Image feature extraction of thyroid nodule based on multiscale fusion[J].Journal of Data Acquisition&Processing,2016,31(5):1004-1009.)
[8]陈海鹏,申铉京,龙建武.采用高斯拟合的全局阈值算法阈值优化框架[J].计算机研究与发展,2016,53(4):892-903.(CHEN Hai-peng,SHEN Xuan-jing,LONG Jian-wu.Threshold optimization framew ork of global thresholding algorithms using Gaussian fitting[J].Journal of Computer Research and Development,2016,53(4):892-903.)
[9]赵亚利,章为川,李云红.图像边缘轮廓自适应阈值的角点检测算法[J].中国图象图形学报,2016,21(11):1502-1514.(ZHAO Ya-li,ZHANG Wei-chuan,LI Yun-hong.Novel contour-based corner detection w ith adaptive threshold[J].Journal of Image and Graphics,2016,21(11):1502-1514.)
[10]刘霖,邓荣钰,刘娟秀,等.基于轮廓曲线的快速高精度可靠图像匹配[J].电子设计工程,2017,25(11):31-34.(LIU Lin,DENG Rong-yu,LIU Juan-xiu,et al.Fast high precision and reliable image matching based on contour curve[J].Electronic Design Engineering,2017,25(11):31-34.)
[2]王丽梅,张宗雪.H型精密运动平台交叉耦合模糊PID同步控制[J].沈阳工业大学学报,2018,40(1):1-5.(WANG Li-mei,ZHANG Zong-xue.Cross-coupling fuzzy PID synchronization control for H precision motion platform[J].Journal of Shenyang University of Technology,2018,40(1):1-5.)
[3]刘小凯,秦川.基于图像块分类阈值优化的改进可逆图像伪装[J].应用科学学报,2018,36(2):237-246.(LIU Xiao-kai,QIN Chuan.Improved reversible image camouflage based on threshold optimization of image block classification[J].Journal of Applied Science,2018,36(2):237-246.)
[4]王文豪,严云洋,姜明新,等.一种去噪声的轮廓提取算法[J].江苏科技大学学报(自然科学版),2017,31(4):519-524.(WANG Wen-hao,YAN Yun-yang,JIANG Mingxin,et al.A noise removing contour extraction algorithm[J].Journal of Jiangsu University of Science and Technology(Natural Science Edition),2017,31(4):519-524.)
[5]韩立博,马安然,任宁宁,等.基于轮廓和灰度特征匹配的X线影像成像部位自动判定方法研究[J].中国医疗器械杂志,2016,40(5):318-321.(HAN Li-bo,MA An-ran,REN Ning-ning,et al.The automatic X-ray radiological positions detection based on contour and gary feature matching[J].Chinese Journal of M edical Lnstrumentation,2016,40(5):318-321.)
[6]张恒,徐刚,刘艳丽,等.基于结构森林的RGB-D图像轮廓提取[J].计算机应用研究,2017,34(2):631-635.(ZHANG Heng,XU Gang,LIU Yan-li,et al.Contour extraction from RGB-D image based on structured forests[J].Application Research of Computers,2017,34(2):631-635.)
[7]王昊,彭博,陈琴,等.基于多尺度融合的甲状腺结节图像特征提取[J].数据采集与处理,2016,31(5):1004-1009.(WANG Hao,PENG Bo,CHEN Qin,et al.Image feature extraction of thyroid nodule based on multiscale fusion[J].Journal of Data Acquisition&Processing,2016,31(5):1004-1009.)
[8]陈海鹏,申铉京,龙建武.采用高斯拟合的全局阈值算法阈值优化框架[J].计算机研究与发展,2016,53(4):892-903.(CHEN Hai-peng,SHEN Xuan-jing,LONG Jian-wu.Threshold optimization framew ork of global thresholding algorithms using Gaussian fitting[J].Journal of Computer Research and Development,2016,53(4):892-903.)
[9]赵亚利,章为川,李云红.图像边缘轮廓自适应阈值的角点检测算法[J].中国图象图形学报,2016,21(11):1502-1514.(ZHAO Ya-li,ZHANG Wei-chuan,LI Yun-hong.Novel contour-based corner detection w ith adaptive threshold[J].Journal of Image and Graphics,2016,21(11):1502-1514.)
[10]刘霖,邓荣钰,刘娟秀,等.基于轮廓曲线的快速高精度可靠图像匹配[J].电子设计工程,2017,25(11):31-34.(LIU Lin,DENG Rong-yu,LIU Juan-xiu,et al.Fast high precision and reliable image matching based on contour curve[J].Electronic Design Engineering,2017,25(11):31-34.)