多特征自适应融合的CamShift算法与ROS跟随小车实现  

Multi-feature Adaptive Fusion CamShift Algorithm for ROS Robot

作　　者：党海鑫 高嵩[1] 曹凯[1] 陈超波[1] 王坤[1] DANG Hai-xin;GAO Song;CAO Kai;CHEN Chao-bo;WANG Kun(School of Electronic Information Engineering,Xi'an Technological University,Xi'an 710021,China)

机构地区：[1]西安工业大学电子信息工程学院,西安710021

出　　处：《科学技术与工程》2021年第10期4141-4147,共7页Science Technology and Engineering

基　　金：陕西省国际科技合作计划(2019KW-014);陕西省重点研发计划(2019GY-072);陕西省教育厅专项科研计划(17JK0369)。

摘　　要：针对传统的具有色彩特征的CamShift算法可以在弱干扰场景中实现良好的跟随,但在光线明显变化或相似颜色干扰的复杂场景中跟踪性能较差的劣势,采用颜色、纹理和边缘特征进行自适应融合的方法来提高算法的抗干扰能力,并通过目标框的尺寸限定和概率密度图的修正进一步提高算法在复杂场景下跟随的稳定性。分别使用机器人操作系统(robot operating system, ROS)跟随小车对优化后的改进算法进行验证,结果证明了该改进算法在光线明显变化或相似颜色干扰的复杂场景下相比于传统的算法具有更高的鲁棒性、跟随精度以及良好的跟随性能。The traditional CamShift algorithm with color features can achieve good follow in weak interference scenes, but has the disadvantage of poor tracking performance in complex scenes with obvious light changes or similar color interference. The method of adaptive fusion of color, texture, and edge features was used to improve the anti-interference ability of the algorithm. The stability of the algorithm in complex scenes was further improved by the size limitation of the target frame and the correction of the probability density map. The robot operating system(ROS) following car was used to verify the improved algorithm. Results show that the improved algorithm has higher robustness, following accuracy, and good following performance than those of the traditional algorithm in complex scenes with obvious light changes or similar color interference.

关 键 词：目标跟随 CAMSHIFT 多特征 自适应融合 机器人操作系统(ROS) 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]