基于头部图像特征的人流计数方法
|
摘要
为提高视频人流智能计数的实时性和准确性,提出一种基于头部图像特征的人流计数方法。根据混合高斯算法检测前景,提取感兴趣的人流区域和相应的边缘;利用基于边缘分类和梯度信息的RHT圆变换识别头部区域,并选取头部面积与头发颜色作为筛选模型参量对头部进行筛选;利用基于Kalman的滤波预测完成跟踪和计数,并将跟踪目标参数反馈给筛选模型,实现更新。实验结果表明,该方法能够有效消除类圆的非头部干扰,提高计数的速度和准确率。
For the purpose of improving the real-time and accuracy of the video intelligent people flow counting,a method for people flow counting based on head image features was proposed.Firstly,the foreground was detected by the mixed Gaussian algorithm to extract the people flow region of interest and the corresponding edges according to the foreground.Then the random Hough transform(RHT)based on edge classification and gradient information was used to recognize the head region and the head filtering model was established to select the targets,where the parameters were determined by the features of head area and hair color.Finally,the tracking and counting were completed by Kalman filter prediction,while the parameters of the tracking target were fed back to the filtering model to realize the update.The experimental results show that this method can effectively eliminate the interference of non-head circle,and improve the speed and accuracy of the people flow counting.
For the purpose of improving the real-time and accuracy of the video intelligent people flow counting,a method for people flow counting based on head image features was proposed.Firstly,the foreground was detected by the mixed Gaussian algorithm to extract the people flow region of interest and the corresponding edges according to the foreground.Then the random Hough transform(RHT)based on edge classification and gradient information was used to recognize the head region and the head filtering model was established to select the targets,where the parameters were determined by the features of head area and hair color.Finally,the tracking and counting were completed by Kalman filter prediction,while the parameters of the tracking target were fed back to the filtering model to realize the update.The experimental results show that this method can effectively eliminate the interference of non-head circle,and improve the speed and accuracy of the people flow counting.
引文
[1]黄凯奇,陈晓棠,康运锋,等.智能视频监控技术综述[J].计算机学报,2015,20(6):1093-1118.HUANG Kaiqi,CHEN Xiaotang,KANG Yunfeng,et al.Intelligent Visual Surveillance:a Review[J].Chinese Journal of Computers,2015,20(6):1093-1118.
[2]刘王胜,冯瑞,一种基于AdaBoost的人脸检测算法[J].计算机工程与应用,2016,52(11):209-214.LIU Wangsheng,FENG Rui.Face Detection Method Based on AdaBoost Algorithm[J].Computer Engineering and Applications,2016,52(11):209-214.
[3]LU J,TAN Y P,WANG G,et al.Image-to-Set Face Recognition Using Locality Repulsion Projections and Sparse Reconstruction-based Similarity Measure[J].IEEE Transactions on Circuits&Systems for Video Technology,2013,23(6):1070-1080.
[4]ZHANG H,CHEN G.Research of Head Detection Based on Curvature Scale Space Algorithm[J].Computer Engineering&Applications,2012,48(14):195-200.
[5]赵军伟,侯清涛,李金屏,等.基于数学形态学和HSI颜色空间的人头检测[J].山东大学学报(工学版),2013,43(2):6-10.ZHAO Junwei,HOU Qingtao,LI Jinping,et al.Head Detection Algorithm Based on Mathematical Morphology and HSI Color Space[J].Journal of Shandong University(Engineering Science),2013,43(2):6-10.
[6]翟卫欣,程承旗.基于Kalman滤波的Camshift运动跟踪算法[J].北京大学学报(自然科学版),2015,51(5):799-804.ZHAI Weixin,CHENG Chengqi.A Camshift Motion Tracking Algorithm Based on Kalman Filter[J].Acta Scientiarum Naturalium Universitatis Pekinensis,2015,51(5):799-804.
[7]CHEN Z,ELLIS T.A Self-adaptive Gaussian Mixture Model[J].Computer Vision&Image Understanding,2014,122(5):35-46.
[8]靳培飞,周莉,刘健,等.基于感兴趣区域提取的行人检测[J].计算机工程与设计,2016,37(11):3035-3039.JIN Peifei,ZHOU Li,LIU Jian,et al.Pedestrian Detection Based on ROI Extraction[J].Computer Engineering and Design,2016,37(11):3035-3039.
[9]CHIU S H,LIAW J J,LIN K H.A Fast Randomized Hough Transform for Circle/Circular Arc Recognition[J].International Journal of Pattern Recognition&Artificial Intelligence,2011,24(3):457-474.
[10]DIRUBERTO C.Generalized Hough Transform for Shape Matching[J].International Journal of Computer Applications,2012,48(1):19-22.
[11]高春霞,董宝田,王爱丽.基于人头颜色空间和轮廓信息的行人检测方法研究[J].交通运输系统工程与信息,2015,15(4):70-77.GAO Chunxia,DONG Baotian,WANG Aili.Pedestrian Detection Method Based on Head Color Space and Outline Information[J].Journal of Transportation Systems Engineering and Information Technology,2015,15(4):70-77.
[12]王炜,郭毓,俞信.基于卡尔曼滤波的多区域关联运动目标跟踪[J].计算机应用,2012,32(11):3174-3177.WANG Wei,GUO Yu,YU Xin.Moving Object Tracking with Related Multi-regions Based on Kalman Filter[J].Journal of Computer Applications,2012,32(11):3174-3177.
[2]刘王胜,冯瑞,一种基于AdaBoost的人脸检测算法[J].计算机工程与应用,2016,52(11):209-214.LIU Wangsheng,FENG Rui.Face Detection Method Based on AdaBoost Algorithm[J].Computer Engineering and Applications,2016,52(11):209-214.
[3]LU J,TAN Y P,WANG G,et al.Image-to-Set Face Recognition Using Locality Repulsion Projections and Sparse Reconstruction-based Similarity Measure[J].IEEE Transactions on Circuits&Systems for Video Technology,2013,23(6):1070-1080.
[4]ZHANG H,CHEN G.Research of Head Detection Based on Curvature Scale Space Algorithm[J].Computer Engineering&Applications,2012,48(14):195-200.
[5]赵军伟,侯清涛,李金屏,等.基于数学形态学和HSI颜色空间的人头检测[J].山东大学学报(工学版),2013,43(2):6-10.ZHAO Junwei,HOU Qingtao,LI Jinping,et al.Head Detection Algorithm Based on Mathematical Morphology and HSI Color Space[J].Journal of Shandong University(Engineering Science),2013,43(2):6-10.
[6]翟卫欣,程承旗.基于Kalman滤波的Camshift运动跟踪算法[J].北京大学学报(自然科学版),2015,51(5):799-804.ZHAI Weixin,CHENG Chengqi.A Camshift Motion Tracking Algorithm Based on Kalman Filter[J].Acta Scientiarum Naturalium Universitatis Pekinensis,2015,51(5):799-804.
[7]CHEN Z,ELLIS T.A Self-adaptive Gaussian Mixture Model[J].Computer Vision&Image Understanding,2014,122(5):35-46.
[8]靳培飞,周莉,刘健,等.基于感兴趣区域提取的行人检测[J].计算机工程与设计,2016,37(11):3035-3039.JIN Peifei,ZHOU Li,LIU Jian,et al.Pedestrian Detection Based on ROI Extraction[J].Computer Engineering and Design,2016,37(11):3035-3039.
[9]CHIU S H,LIAW J J,LIN K H.A Fast Randomized Hough Transform for Circle/Circular Arc Recognition[J].International Journal of Pattern Recognition&Artificial Intelligence,2011,24(3):457-474.
[10]DIRUBERTO C.Generalized Hough Transform for Shape Matching[J].International Journal of Computer Applications,2012,48(1):19-22.
[11]高春霞,董宝田,王爱丽.基于人头颜色空间和轮廓信息的行人检测方法研究[J].交通运输系统工程与信息,2015,15(4):70-77.GAO Chunxia,DONG Baotian,WANG Aili.Pedestrian Detection Method Based on Head Color Space and Outline Information[J].Journal of Transportation Systems Engineering and Information Technology,2015,15(4):70-77.
[12]王炜,郭毓,俞信.基于卡尔曼滤波的多区域关联运动目标跟踪[J].计算机应用,2012,32(11):3174-3177.WANG Wei,GUO Yu,YU Xin.Moving Object Tracking with Related Multi-regions Based on Kalman Filter[J].Journal of Computer Applications,2012,32(11):3174-3177.