基于局部灰度熵的人体检测方法

Detection of human body based on local gray entropy

针对造成低对比度环境下运动人体检测困难的两个主要因素:拍摄时光线昏暗和拍摄时距离较远,引入局部灰度熵概念,根据局部灰度熵可以准确地反映样本的离散程度且与样本的灰度均值无关这一原理,提出基于局部灰度熵的人体目标检测算法。建立背景模型,运用泰勒展开式简化局部灰度熵计算公式,计算邻域窗口内运动物体与背景模型的局部灰度熵值之差,通过检测率与虚警率对算法进行的评价,得到两种低对比度情况下可以获取运动人体目标的局部灰度熵差值的最佳阈值。实验结果表明,在低对比度环境下,基于局部灰度熵的人体检测算法能够有效地检测出运动人体目标。

With regard to the two main factors that cause the difficulty of human-detection under the dim contrast environment,a concept of the local gray entropy was introduced,and then an algorithm of human target detection based on local gray entropy was proposed for the reason that the local gray entropy could reflect the discrete level accurately,and it was independent on the average gray.After the background model was established,the local gray entropy difference between the moving objects and background model inside the domain windows was calculated by using calculation formula of local gray entropy simplified through Taylor expansion.And the ratio of detection and the ratio of false-alarm of the algorithm were evaluated.The optimal thresholds on the differential of the local gray entropy which could get the human body under the two conditions of low contrast were obtained.The experimental results show that the algorithm of human target detection based on the trait of the local gray entropy can obtain the moving human targets effectively under the dim contrast environment.