基于机器学习识别与标记分水岭分割的盲道图像定位
|
摘要
视觉导盲仪是一种旨在解决盲人出行困难的导盲设备,为了实现视觉导盲仪诱导盲人找到盲道并沿盲道行走,提出了一种基于机器学习识别与标记分水岭分割的盲道图像定位算法,通过离线训练与在线的识别、分割来定位盲道区域。首先对盲道图像进行视角变换的预处理,根据盲道的地面方程将变化的倾斜视角转换为固定的俯视视角,消除射影变换带来的失真;然后利用局部二进制模式描述子提取鸟瞰图的纹理特征,以自适应增强学习算法离线训练盲道识别分类器;进而利用分类器对鸟瞰图像进行在线识别,粗略确定盲道区域;最后将识别结果进行形态学处理后作为标记,利用标记分水岭算法得到精确分割的盲道区域并定位盲道中心线。在研制的视觉导盲仪上进行验证,结果表明盲道定位准确率达到了95.44%,速度平均每秒8帧,具有高准确率的同时达到实时性要求,为盲道的准确三维定位提供了必要条件。
VTA(Visual Travel Aids)are devices used for addressing traveling difficulties of visually impaired people.To develop VTA for guiding visually impaired people to blind sidewalks,a method for blind sidewalk image location was presented based on machine learning recognition and a marked watershed algorithm.The algorithm located blind sidewalks by combining offline training with online recognition and segmentation.First,a blind sidewalk image was pretreated by converting an original image from a camera gradient view into an aerial view based on the plane equation of the blind sidewalk.The pretreating eliminates distortions.A Local Binary Pattern(LBP)descriptor then extracted the texture features of the aerial-view images.An offline cascade classifier trained through Adaboost recognized the blind sidewalk based on the LBP descriptor.The cascade classifier recognized the aerialview image online and roughly identified the blind area.The recognition results were then morphologically processed as markers to obtain the exact segmentation of the blind area through a marked watershed algorithm.Finally,the segmentation results were used to locate the centerline of the blind sidewalk.The algorithm was validated on the VTA.The experimental result showed that the blind sidewalk localization achieved 95.44%accuracy with an average speed of 8 frame/s.It exhibited a high accuracy rate while satisfying the real-time requirement,which are the necessary conditions for accurate3 Dlocalization of blind sidewalks.
VTA(Visual Travel Aids)are devices used for addressing traveling difficulties of visually impaired people.To develop VTA for guiding visually impaired people to blind sidewalks,a method for blind sidewalk image location was presented based on machine learning recognition and a marked watershed algorithm.The algorithm located blind sidewalks by combining offline training with online recognition and segmentation.First,a blind sidewalk image was pretreated by converting an original image from a camera gradient view into an aerial view based on the plane equation of the blind sidewalk.The pretreating eliminates distortions.A Local Binary Pattern(LBP)descriptor then extracted the texture features of the aerial-view images.An offline cascade classifier trained through Adaboost recognized the blind sidewalk based on the LBP descriptor.The cascade classifier recognized the aerialview image online and roughly identified the blind area.The recognition results were then morphologically processed as markers to obtain the exact segmentation of the blind area through a marked watershed algorithm.Finally,the segmentation results were used to locate the centerline of the blind sidewalk.The algorithm was validated on the VTA.The experimental result showed that the blind sidewalk localization achieved 95.44%accuracy with an average speed of 8 frame/s.It exhibited a high accuracy rate while satisfying the real-time requirement,which are the necessary conditions for accurate3 Dlocalization of blind sidewalks.
引文
[1]MEHTA U,ALIM M,KUMAR S.Smart path guidance mobile aid for visually disabled persons[J].Procedia Computer Science,2017,105:52-56.
[2] GONNOT T,SANIIE J.Integrated machine vision and communication system for blind navigation and guidance[C].IEEE International Conference on Electro Information Technology.IEEE,2016:0187-0191.
[3]张克华,王书平,尹晓红,等.复杂室内环境下的单一Kinect导盲系统[J].光学精密工程,2015,23(8):2419-2427.ZHANG K H,WANG SH P,YIN X H,et al..Blind navigation aid system with Kinect sensor in complex indoor environments[J].Guangxue Jingmi Gongcheng/optics&Precision Engineering,2015,23(8):2419-2427.(in Chinese)
[4] BOLOGNA G,DEVILLE B,PUN T.Blind Navigation along a Sinuous Path by Means of the See ColOr Interface[C].International Work-Conference on the Interplay Between Natural and Artificial Computation.Springer-Verlag,2009:235-243.
[5]闫志杰,汪剑鸣,窦汝振,等.基于颜色聚类和直线检测的自适应盲道区域分割算法[J].天津工业大学学报,2010,29(1):80-84.YAN ZH J,WANG J M,DOU R ZH,et al..Sidewalk for the blind adaptive segmentation based on color clustering and line detection[J].Journal of Tianjin Polytechnic University,2010,29(1):80-84.(in Chinese)
[6] TANG J,JING X,HE D,et al..Blind-road location and recognition in natural scene[C].Wri World Congress on Computer Science and Information Engineering.IEEE Computer Society,2009:283-287.
[7]彭玉青,薛杰,郭永芳.基于颜色纹理信息的盲道识别算法[J].计算机应用,2014,34(12):3585-3588.PENG Y J,XUE J,GUO Y F.Blind road recognition algorithm based on color and texture information[J].Electronic Information Warfare Technology,2014.(in Chinese)
[8]王民,肖磊,杨放.基于显著性检测和改进投影字典对的盲道分割[J].激光与光电子学进展,2017(4):133-140.WANG M,XIAO L,YANG F.Blind road segmentation based on saliency detection and improved projective dictionary pair[J].Laser&Optoelectronics Progress,2017,54(4):041001.(in Chinese)
[9] FENG J X,WEI T.Wearable guide system for obstacle avoidance based on stereo computer vision[C].International Conference on Manufacturing Science and Engineering,2015:1811-1817.
[10] RICHARD HARTLEY ANDREW ZISSERMAN.计算机视觉中的多视图几何[M].安徽大学出版社,2002.HARTLEY R,ZISSERMAN A.Multiple View Geometry in Computer Vision[M].Anhui:Anhui University Press,2003.(in Chinese)
[11]张静,叶玉堂,谢煜,等.金属圆柱工件缺陷的光电检测[J].光学精密工程,2014,22(7):1871-1876.ZHANG J,YE Y T,XIE Y,et al..Optoelectronic inspection of defects for metal cylindrical workpieces[J].Opt.Precision Eng.,2014,22(7):1871-1876.(in Chinese)
[12] SINGH G,CHHABRA I.Effective and fast face recognition system using complementary OC-LBP and HOG feature descriptors with SVM classifier[J].Journal of Information Technology Research,2018,11(1):91-110.
[13] REN J,JIANG X,YUAN J.LBP-structure optimization with symmetry and uniformity regularizations for scene classification[J].IEEE Signal Processing Letters,2017,24(1):37-41.
[14]武明虎,宋冉冉,刘敏.结合HSV与纹理特征的视频阴影消除算法[J].中国图象图形学报,2017,22(10):1373-1380.WU M H,SONG R R,LIU M.Video shadow elimination algorithm by combining HSV with texture features[J].Journal of Image&Graphics,2017,22(10):1373-1380.(in Chinese)
[15]黄文博,王珂,燕杨.彩色视网膜眼底图像血管自动检测方法[J].光学精密工程,2017,25(5):1378-1386.HUANG W B,WANG K,YAN Y.Automatic detection method of blood vessel for color retina fundus images[J].Opt.Precision Eng.,2017,25(5):1378-1386.(in Chinese)
[16]曹莹,苗启广,刘家辰,等.AdaBoost算法研究进展与展望[J].自动化学报,2013,39(6):745-758.CAO Y,MIAO Q G,LIU J CH,et al..Advance and prospects of adaBoost algorithm[J].Acta Automatica Sinica,2013,39(6):745-758.(in Chinese)
[17] RAJESHWARI J,KARIBASAPPA K.Adaboost modular tensor locality preservative projection:face detection in video using Adaboost modularbased tensor locality preservative projections[J].Iet Computer Vision,2017,10(7):670-678.
[18] WALSE K H,DHARASKAR R V,THAKARE V M.A study of human activity recognition using adaboost classifiers on WISDM dataset[J].Iioab Journal,2016,7(2):68-76.
[19]徐超,高敏,杨锁昌,等.视觉注意机制下的粒子窗快速目标检测[J].光学精密工程,2015,23(11):3227-3237.XU CH,GAO M,YANG S CH,et al..Visual attention mechanism-aided fast target detection by particle window[J].Opt.Precision Eng.,2015,23(11):3227-3237.(in Chinese)
[2] GONNOT T,SANIIE J.Integrated machine vision and communication system for blind navigation and guidance[C].IEEE International Conference on Electro Information Technology.IEEE,2016:0187-0191.
[3]张克华,王书平,尹晓红,等.复杂室内环境下的单一Kinect导盲系统[J].光学精密工程,2015,23(8):2419-2427.ZHANG K H,WANG SH P,YIN X H,et al..Blind navigation aid system with Kinect sensor in complex indoor environments[J].Guangxue Jingmi Gongcheng/optics&Precision Engineering,2015,23(8):2419-2427.(in Chinese)
[4] BOLOGNA G,DEVILLE B,PUN T.Blind Navigation along a Sinuous Path by Means of the See ColOr Interface[C].International Work-Conference on the Interplay Between Natural and Artificial Computation.Springer-Verlag,2009:235-243.
[5]闫志杰,汪剑鸣,窦汝振,等.基于颜色聚类和直线检测的自适应盲道区域分割算法[J].天津工业大学学报,2010,29(1):80-84.YAN ZH J,WANG J M,DOU R ZH,et al..Sidewalk for the blind adaptive segmentation based on color clustering and line detection[J].Journal of Tianjin Polytechnic University,2010,29(1):80-84.(in Chinese)
[6] TANG J,JING X,HE D,et al..Blind-road location and recognition in natural scene[C].Wri World Congress on Computer Science and Information Engineering.IEEE Computer Society,2009:283-287.
[7]彭玉青,薛杰,郭永芳.基于颜色纹理信息的盲道识别算法[J].计算机应用,2014,34(12):3585-3588.PENG Y J,XUE J,GUO Y F.Blind road recognition algorithm based on color and texture information[J].Electronic Information Warfare Technology,2014.(in Chinese)
[8]王民,肖磊,杨放.基于显著性检测和改进投影字典对的盲道分割[J].激光与光电子学进展,2017(4):133-140.WANG M,XIAO L,YANG F.Blind road segmentation based on saliency detection and improved projective dictionary pair[J].Laser&Optoelectronics Progress,2017,54(4):041001.(in Chinese)
[9] FENG J X,WEI T.Wearable guide system for obstacle avoidance based on stereo computer vision[C].International Conference on Manufacturing Science and Engineering,2015:1811-1817.
[10] RICHARD HARTLEY ANDREW ZISSERMAN.计算机视觉中的多视图几何[M].安徽大学出版社,2002.HARTLEY R,ZISSERMAN A.Multiple View Geometry in Computer Vision[M].Anhui:Anhui University Press,2003.(in Chinese)
[11]张静,叶玉堂,谢煜,等.金属圆柱工件缺陷的光电检测[J].光学精密工程,2014,22(7):1871-1876.ZHANG J,YE Y T,XIE Y,et al..Optoelectronic inspection of defects for metal cylindrical workpieces[J].Opt.Precision Eng.,2014,22(7):1871-1876.(in Chinese)
[12] SINGH G,CHHABRA I.Effective and fast face recognition system using complementary OC-LBP and HOG feature descriptors with SVM classifier[J].Journal of Information Technology Research,2018,11(1):91-110.
[13] REN J,JIANG X,YUAN J.LBP-structure optimization with symmetry and uniformity regularizations for scene classification[J].IEEE Signal Processing Letters,2017,24(1):37-41.
[14]武明虎,宋冉冉,刘敏.结合HSV与纹理特征的视频阴影消除算法[J].中国图象图形学报,2017,22(10):1373-1380.WU M H,SONG R R,LIU M.Video shadow elimination algorithm by combining HSV with texture features[J].Journal of Image&Graphics,2017,22(10):1373-1380.(in Chinese)
[15]黄文博,王珂,燕杨.彩色视网膜眼底图像血管自动检测方法[J].光学精密工程,2017,25(5):1378-1386.HUANG W B,WANG K,YAN Y.Automatic detection method of blood vessel for color retina fundus images[J].Opt.Precision Eng.,2017,25(5):1378-1386.(in Chinese)
[16]曹莹,苗启广,刘家辰,等.AdaBoost算法研究进展与展望[J].自动化学报,2013,39(6):745-758.CAO Y,MIAO Q G,LIU J CH,et al..Advance and prospects of adaBoost algorithm[J].Acta Automatica Sinica,2013,39(6):745-758.(in Chinese)
[17] RAJESHWARI J,KARIBASAPPA K.Adaboost modular tensor locality preservative projection:face detection in video using Adaboost modularbased tensor locality preservative projections[J].Iet Computer Vision,2017,10(7):670-678.
[18] WALSE K H,DHARASKAR R V,THAKARE V M.A study of human activity recognition using adaboost classifiers on WISDM dataset[J].Iioab Journal,2016,7(2):68-76.
[19]徐超,高敏,杨锁昌,等.视觉注意机制下的粒子窗快速目标检测[J].光学精密工程,2015,23(11):3227-3237.XU CH,GAO M,YANG S CH,et al..Visual attention mechanism-aided fast target detection by particle window[J].Opt.Precision Eng.,2015,23(11):3227-3237.(in Chinese)