多自由度人体运动动态图像目标检测方法仿真  

Simulation of Object Detection Method in Multi Degree of Freedom Human Motion Dynamic Image

作　　者：唐开悦 王思涵 TANG Kai-yue;WANG Si-han(Guangxi Normal University,Guangxi Guilin 541006,China)

机构地区：[1]广西师范大学,广西桂林541006

出　　处：《计算机仿真》2021年第9期199-202,437,共5页Computer Simulation

摘　　要：采用当前方法对人体运动动态图像进行目标检测时,无法消除图像中存在的阴影区域,导致图像清晰度低、位置误差大和检测效率低。提出多自由度人体运动动态图像目标检测方法,构建高斯背景模型,分离图像的背景区域和目标区域,将目标区域变换到HSI空间中,结合小区域去除方法和数学形态学处理,根据阴影区域饱和度高和亮度值低的特点确定目标区域中存在的阴影,通过匹配补偿去除目标区域中的阴影部分。根据目标区域的灰度分布构建目标更新模型,实现多自由度人体运动动态图像目标的检测。实验结果表明,所提方法的图像清晰度高、位置误差小、检测效率高。When the current method is used for target detection of human motion dynamic image, it can not eliminate the shadow area in the image, resulting in low image definition, large position error and low detection efficiency.In this regard, this paper designed a multi-degree of freedom detection method for human motion dynamic image.The background region and the target region of the image were separated via the establishment of Gaussian background model.The target region was transformed into HSI space.Small area removal method and mathematical morphology were introduced, and based on the high saturation and low brightness of the shadow area, the shadow in the target area was determined.According to the matching compensation principle, the shadow in the target area was eliminated.The gray distribution of the target region was applied to build the target update model, achieving the multi degree of freedom of human motion dynamic image target detection.The experimental results show that the method has high image definition, detection efficiency and minor position error.

关 键 词：人体运动动态图像 背景建模 阴影消除方法 目标检测方法 灰度分布 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]