摘要
为了进一步提高跟踪系统中目标检测的精准性,适应不同的环境变化,对检测阈值的选取和图像的去噪提出了新的算法。结合似然函数和贝叶斯判别准则理论,根据信号和噪声的统计分布规律,得到一个自适应的动态阈值。利用该阈值对二值化图像的水平投影进行优化处理,去除噪声,得到待检测的目标。实验结果表明,该方法克服了目标完整性和抑制噪声之间的矛盾,可以有效地改善光线变化带来的噪声影响,在不同的场景情况下都可以得到满意的检测结果。
In order to increase the accuracy of object detecting,a new method is proposed to select the threshold and clean out the noise when the moving objects are detected.Combined with the Bayesian rules and likelihood function,a dynamic and adaptive threshold is obtained.The level projection is processed based on the threshold and the noise is removed,the objects awaiting detection are obtained.The experimental result shows that the proposed method can overcome the contradiction between the de-noising and integrity of the objects,and it can improve the effect of the light change and obtain a satisfactory detected result.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2011年第4期933-937,共5页
Systems Engineering and Electronics
基金
国家自然科学基金(60775023
60975025)
教育部留学回国人员科研启动基金资助课题
关键词
目标检测
阈值选取
贝叶斯准则
水平投影
动态阈值
object detection
threshold selection
Bayesian rules
horizontal projection
dynamic threshold
