基于HSV与RGB颜色空间的人体检测与跟踪算法  

Human Detection and Tracking Algorithm Based on HSV and RGB Color Space

作　　者：杨鸽[1] 郑嘉龙 王莹 YANG Ge;ZHENG Jia-long;WANG Ying(SichuanWater Conservancy Vocational College,Chengdu 611231 China)

机构地区：[1]四川水利职业技术学院,四川成都611231

出　　处：《自动化技术与应用》2022年第9期17-20,28,共5页Techniques of Automation and Applications

基　　金：四川水利职业技术学院院级科研项目(KY2019-13)。

摘　　要：针对在光照变化、人影干扰等特殊场景下人体灰度与环境背景灰度相近时,难以将人体与背景分割开来这一问题,提出了一种基于HSV与RGB颜色空间的人体检测与跟踪算法。首先,在RGB颜色空间构建高斯背景更新模型;然后,使用背景减除法计算待检测图像与背景图像R、G、B分量的欧氏距离检测出人体目标,根据检测结果实时更新背景模型;最后,根据阴影在HSV空间下的特点对阴影进行检测与去除,计算目标区域颜色直方图的相关性对目标进行跟踪。实验结果表明,在复杂场景下此方法能够有效的消除因光照和人体形变等因素造成的干扰,准确实现人体检测和跟踪,算法具有较好的鲁棒性。It is difficult to present the difference between the body and the background when the human body gray is similar to the environment background gray because of illumination change or shadow interference or others.To solve this problem,it comes up with a kind of human detection and tracking algorithm based on HSV and RGB color space. First of all, Its background updating method is set in color space by adopting the adaptive Gaussian mixture model. Then use the background subtraction division to calculate the Euclidean distance of corresponding points of the R.G. B component of the image and the background image to be detected with the purpose of segmenting the foreground target. Finally, according to the characteristics of the shadows under the HSV space for shadow detection and removal, it adopts the method of calculating target area for the color histogram correlation to track the target. The experiment results show that in complex scenarios, this proposed algorithm can validly eliminate the interference due to factors such as light and human body deformation and accurately realize the human detection and tracking. Besides,the algorithm provides tremendous robustness.

关 键 词：人体检测 背景差分 HSV 阴影检测 目标跟踪 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP183[自动化与计算机技术—计算机科学与技术]