红外图像目标跟踪在智能网联汽车的应用研究
|
摘要
过去采用数据融合方法进行的汽车目标跟踪,未对目标图像信息有效处理,图像噪声影响较强,无法准确的对目标实施识别和跟踪,研究红外图像目标跟踪在智能网联汽车的应用过程,采用红外图像成像原理获取智能网联汽车的目标红外图像,通过滤波降噪、图像增强和图像分割操作降低红外图像的噪声、增强目标与背景对比度,实现目标与红外图像背景的有效分割;在此基础上,采用均值漂移目标跟踪算法,在新的红外图像中确定目标待选位置,并确保描述目标直方图与备选目标直方图概率分布相似度的巴氏系数最大,实现智能网联汽车对目标的准确跟踪。实验结果说明,红外图像目标跟踪在提升智能网联汽车目标识别结果上有显著作用,识别简单场景和复杂场景目标的正确率分别为0. 966和0. 565,都高于数据融合方法,且目标跟踪效率高。
In the past,the method of data fusion was used to track car targets. It did not effectively process target image information,and the image noise had a strong influence. It was impossible to accurately identify and track targets,and the application process of infrared image target tracking in intelligent network vehicles was studied. The infrared image imaging principle is used to obtain the infrared image of the intelligent network car. The noise reduction,image enhancement and image segmentation operation of the filter wave reduce the noise of the infrared image,enhance the contrast between the target and the background,and achieve the effective segmentation of the target and the infrared image background. On this basis,using the mean drift target tracking algorithm,the target's location is determined in the new infrared image,and the Barthel coefficient describing the probability distribution similarity between the target histogram and the alternative target histogram is ensured to realize the accurate tracking of the targets of the intelligent network car. The experimental results show that infrared image target tracking plays a significant role in improving the target recognition results of intelligent network-connected vehicles. The accuracy rates of simple scene recognition and complex scene target reach up to 0.966 and 0.565,respectively,which are higher than the data fusion method and the target tracking efficiency.
In the past,the method of data fusion was used to track car targets. It did not effectively process target image information,and the image noise had a strong influence. It was impossible to accurately identify and track targets,and the application process of infrared image target tracking in intelligent network vehicles was studied. The infrared image imaging principle is used to obtain the infrared image of the intelligent network car. The noise reduction,image enhancement and image segmentation operation of the filter wave reduce the noise of the infrared image,enhance the contrast between the target and the background,and achieve the effective segmentation of the target and the infrared image background. On this basis,using the mean drift target tracking algorithm,the target's location is determined in the new infrared image,and the Barthel coefficient describing the probability distribution similarity between the target histogram and the alternative target histogram is ensured to realize the accurate tracking of the targets of the intelligent network car. The experimental results show that infrared image target tracking plays a significant role in improving the target recognition results of intelligent network-connected vehicles. The accuracy rates of simple scene recognition and complex scene target reach up to 0.966 and 0.565,respectively,which are higher than the data fusion method and the target tracking efficiency.
引文
[1]王肖,李克强,王建强,等.基于三维激光雷达的智能车辆目标参数辨识[J].汽车工程,2016,38(9):1146-1152.
[2]何欣龙,刘文浩,王继芬.红外光谱结合多元统计学检验汽车前保险杠[J].光散射学报,2018,30(1):70-76.
[3]蒲静松,许东,刘乙君,等.星空背景下红外弱小目标的快速检测[J].激光与红外,2017,47(4):513-518.
[4]刘帆,刘鹏远,李兵,等.Tensor Flow平台下的视频目标跟踪深度学习模型设计[J].激光与光电子学进展,2017,54(9):277-285.
[5]张建伟.基于区域特征匹配的扩展目标高精度跟踪[J].激光与光电子学进展,2015,52(5):79-86.
[6]岳元,彭亮节.SDI红外图像在火炮光电跟踪系统中的应用[J].激光与红外,2016,46(8):1023-1027.
[7]苑晶,刘钢墩,孙沁璇.激光与单目视觉融合的移动机器人运动目标跟踪[J].控制理论与应用,2016,33(2):196-204.
[8]佚名.基于核相关滤波器的多目标跟踪算法[J].激光与光电子学进展,2018,55(9).
[9]巩学美,高昆,王研,等.一种基于概率Hough变换的遥感图像中线目标检测新方法[J].影像科学与光化学,2017,35(2):162-167.
[10]ZHAO X,HE Z,ZHANG S,et al.Robust pedestrian detection in thermal infrared imagery using a shape distribution histogram feature and modified sparse representation classification[J].Pattern Recognition,2015,48(6):1947-1960.
[11]于洋,李丹.跟踪图像生成技术应用于光电设备模拟训练[J].液晶与显示,2015,30(2):347-352.
[12]SHIEH W Y,HSU C C J,WANG T H.Vehicle Positioning and Trajectory Tracking by Infrared Signal-Direction Discrimination for Short-Range Vehicle-to-Infrastructure Communication Systems[J].IEEE Transactions on Intelligent Transportation Systems,2018,PP(99):1-12.
[13]刘程,阳洪.自适应计步检测算法研究[J].压电与声光,2015,37(2):258-261.
[14]杨福才,杨德东,毛宁,等.基于稀疏编码直方图的稳健红外目标跟踪[J].光学学报,2017,37(11):196-206.
[15]田岳鑫,高昆,庄幽文,等.一种基于最优非线性滤波的红外弱目标检测前跟踪算法研究[J].光学技术,2015,41(4):369-372.
[16]王玲玲,王宏.基于激光传感器的自主循迹智能车设计[J].电子测量技术,2017,40(5):193-196.
[17]粟培,王琛,郭军,等.高速运动目标追踪摄像及识别跟踪算法研究[J].光通信技术,2015,39(3).
[2]何欣龙,刘文浩,王继芬.红外光谱结合多元统计学检验汽车前保险杠[J].光散射学报,2018,30(1):70-76.
[3]蒲静松,许东,刘乙君,等.星空背景下红外弱小目标的快速检测[J].激光与红外,2017,47(4):513-518.
[4]刘帆,刘鹏远,李兵,等.Tensor Flow平台下的视频目标跟踪深度学习模型设计[J].激光与光电子学进展,2017,54(9):277-285.
[5]张建伟.基于区域特征匹配的扩展目标高精度跟踪[J].激光与光电子学进展,2015,52(5):79-86.
[6]岳元,彭亮节.SDI红外图像在火炮光电跟踪系统中的应用[J].激光与红外,2016,46(8):1023-1027.
[7]苑晶,刘钢墩,孙沁璇.激光与单目视觉融合的移动机器人运动目标跟踪[J].控制理论与应用,2016,33(2):196-204.
[8]佚名.基于核相关滤波器的多目标跟踪算法[J].激光与光电子学进展,2018,55(9).
[9]巩学美,高昆,王研,等.一种基于概率Hough变换的遥感图像中线目标检测新方法[J].影像科学与光化学,2017,35(2):162-167.
[10]ZHAO X,HE Z,ZHANG S,et al.Robust pedestrian detection in thermal infrared imagery using a shape distribution histogram feature and modified sparse representation classification[J].Pattern Recognition,2015,48(6):1947-1960.
[11]于洋,李丹.跟踪图像生成技术应用于光电设备模拟训练[J].液晶与显示,2015,30(2):347-352.
[12]SHIEH W Y,HSU C C J,WANG T H.Vehicle Positioning and Trajectory Tracking by Infrared Signal-Direction Discrimination for Short-Range Vehicle-to-Infrastructure Communication Systems[J].IEEE Transactions on Intelligent Transportation Systems,2018,PP(99):1-12.
[13]刘程,阳洪.自适应计步检测算法研究[J].压电与声光,2015,37(2):258-261.
[14]杨福才,杨德东,毛宁,等.基于稀疏编码直方图的稳健红外目标跟踪[J].光学学报,2017,37(11):196-206.
[15]田岳鑫,高昆,庄幽文,等.一种基于最优非线性滤波的红外弱目标检测前跟踪算法研究[J].光学技术,2015,41(4):369-372.
[16]王玲玲,王宏.基于激光传感器的自主循迹智能车设计[J].电子测量技术,2017,40(5):193-196.
[17]粟培,王琛,郭军,等.高速运动目标追踪摄像及识别跟踪算法研究[J].光通信技术,2015,39(3).