基于高速视觉目标跟踪系统的自动调焦算法的设计
|
摘要
为了解决高速视觉目标跟踪系统在进行目标跟踪过程中目标的尺度与清晰度的变化,提出一种基于尺度不变性的自动调焦算法。该算法包括基于改进型二维OTSU分割的自动变焦算法和基于二维图像信息熵的对焦算法,通过高速视觉目标跟踪系统获得目标区域,计算目标像素面积和清晰度评价值,转化为可调焦镜头变焦步数和对焦步数后通过FPGA发送调焦指令,最终完成自动调焦动作。设计了自动调焦验证系统以验证算法的性能并得出满足追踪要求的最佳调焦参数。实验证明,该算法在高速目标跟踪系统的跟踪过程中可以较好地完成自动调焦任务。
For the purpose of resolving the change of objects' scale and clarity in the process of the high speed object tracking system,an auto-focusing algorithm based on scale invariance was proposed.It includes an improved 2 D Otsu segmentation algorithm for automatic zooming and a tuning focus algorithm based on 2 Dimage information entropy,it obtains the region of object through the high speed vision object tracking system,calculates the object pixel area and the sharpness evaluation value,and it finally completes the auto focus operation by sending the transformed zooming steps and tuning focusing steps.An automatic focusing verification system was designed to verify the performance of the algorithm and derive the optimal focusing parameters to meet tracking.Experiments prove that this algorithm can complete the automatic focusing task in the tracking process of high-speed object tracking system.
For the purpose of resolving the change of objects' scale and clarity in the process of the high speed object tracking system,an auto-focusing algorithm based on scale invariance was proposed.It includes an improved 2 D Otsu segmentation algorithm for automatic zooming and a tuning focus algorithm based on 2 Dimage information entropy,it obtains the region of object through the high speed vision object tracking system,calculates the object pixel area and the sharpness evaluation value,and it finally completes the auto focus operation by sending the transformed zooming steps and tuning focusing steps.An automatic focusing verification system was designed to verify the performance of the algorithm and derive the optimal focusing parameters to meet tracking.Experiments prove that this algorithm can complete the automatic focusing task in the tracking process of high-speed object tracking system.
引文
[1]Yang Y X,Yang J,Zhang Z X,et al.High-speed visual target tracking with mixed rotation invariant description and skipping searching[J].Sci.China Inf.Sci.,2017,60(6):062401.
[2]Ishii I,Taniguchi T,Yamamoto K,et al.1 000fps real-time optical flow detection system[J].Proc.of SPIE,2010,7538:75380M.
[3]尹宏鹏,陈波,柴毅,等.基于视觉的目标检测与跟踪综述[J].自动化学报,2016,42(10):1466-1489.Yin Hongpeng,Chen Bo,Chai Yi,et al.Vision-based target detection and tracking review[J].Act.Automatica Sin.,2016,42(10):1466-1489.
[4]Moeslund T B,Granum E.A survey of computer vision-based human motion capture[J].Computer Vision and Image Understanding,2001,81(3):231-268.
[5]Lepetit V,Fua P.Monocular model-based 3Dtracking of rigid objects:A survey[J].Foundations and Trendsin Computer Graphics and Vision,2005,1(1):1-89.
[6]Zhang J,Hu J.Image segmentation based on 2DOtsu method with histogram analysis[C]//2008Inter.Conf.on Computer Science and Software Engin.,2008,6:105-108.
[7]王坤,张杨.改进二维OTSU和自适应遗传算法的红外图像分割[J].系统仿真学报,2017,29(6):1229-1236.Wang Kun,Zhang Yang.Improved two-dimensional OTSU and adaptive genetic algorithm for infrared image segmentation[J].J.of System Simulation,2017,29(6):1229-1236.
[8]王凤朝,黄树采,韩朝超.基于改进的二维OTSU法的图像分割法[J].航空计算技术,2008,38(4):4-7.Wang Fengchao,Huang Shucai,Han Chaochao.Image segmentation method based on improved 2-D OTSU’s method[J].Aviation Computing Technol.,2008,38(4):4-7.
[9]Abutaleb A S.Automatic thresholding of gray-level pictures using two-dimensional entropy[J].Computer Vision,Graphics,and Image Proc.,1989,47(1):22-32.
[10]Wohl Ishay,Zurgil Naomi,Hakuk Yaron,et al.Fluctuation of information entropy measures in cell image[J].Entropy,2017,19(10):565.
[2]Ishii I,Taniguchi T,Yamamoto K,et al.1 000fps real-time optical flow detection system[J].Proc.of SPIE,2010,7538:75380M.
[3]尹宏鹏,陈波,柴毅,等.基于视觉的目标检测与跟踪综述[J].自动化学报,2016,42(10):1466-1489.Yin Hongpeng,Chen Bo,Chai Yi,et al.Vision-based target detection and tracking review[J].Act.Automatica Sin.,2016,42(10):1466-1489.
[4]Moeslund T B,Granum E.A survey of computer vision-based human motion capture[J].Computer Vision and Image Understanding,2001,81(3):231-268.
[5]Lepetit V,Fua P.Monocular model-based 3Dtracking of rigid objects:A survey[J].Foundations and Trendsin Computer Graphics and Vision,2005,1(1):1-89.
[6]Zhang J,Hu J.Image segmentation based on 2DOtsu method with histogram analysis[C]//2008Inter.Conf.on Computer Science and Software Engin.,2008,6:105-108.
[7]王坤,张杨.改进二维OTSU和自适应遗传算法的红外图像分割[J].系统仿真学报,2017,29(6):1229-1236.Wang Kun,Zhang Yang.Improved two-dimensional OTSU and adaptive genetic algorithm for infrared image segmentation[J].J.of System Simulation,2017,29(6):1229-1236.
[8]王凤朝,黄树采,韩朝超.基于改进的二维OTSU法的图像分割法[J].航空计算技术,2008,38(4):4-7.Wang Fengchao,Huang Shucai,Han Chaochao.Image segmentation method based on improved 2-D OTSU’s method[J].Aviation Computing Technol.,2008,38(4):4-7.
[9]Abutaleb A S.Automatic thresholding of gray-level pictures using two-dimensional entropy[J].Computer Vision,Graphics,and Image Proc.,1989,47(1):22-32.
[10]Wohl Ishay,Zurgil Naomi,Hakuk Yaron,et al.Fluctuation of information entropy measures in cell image[J].Entropy,2017,19(10):565.