摘要
由于雾天、阴天、高噪环境等外部环境因素的影响,常常出现跟踪过程中易丢失目标的现象。该文通过改进多尺度Retinex算法增强图像,在压缩图像中高亮区域范围的同时,拉伸暗区的动态范围,使得暗区的细节信息明显提高,并结合Camshift算法,使用预测位置误差方法估计运动目标搜索范围,并使用滤波器对参数修正,有效克服了Camshift算法自身的缺陷。
Due to the fog, cloudy, high noise environment and other external factors, it is easy to lose target in the tracking pro- cess. In this paper, through the improved multi-scale Retinex algorithm, the highlighted area in the image is compressed while the dynamic range of the dark region is stretched. The details of the dark area increased significantly. Combined with Camshift algorithm, the predictive position error method was used to estimate the range of motion target. By using the filter to modify the parameters, effectively overcome the shortcomings of the Camshift algorithm itself.
作者
龚俐
厉丹
郁琳玲
黄云
张英
GONG Li, LI Dan, YU Lin-ling, HUANG Yun, ZHANG Ying （College of Information and Electrical Engineering, Xuzhou Institute of Technology, Xuzhou 221008, China）
出处
《电脑知识与技术》
2016年第12期216-218,共3页
Computer Knowledge and Technology
基金
徐州工程学院2016年大学生创新创业训练计划项目(xcx2016080)
