摘要
运动对象检测是计算机视觉应用中的一个重要问题。提出了一种新的检测运动对象的算法。首先通过计算图像块的小波变换域不同频率的纹理特征从而得到图像的特征,然后利用前景图像和背景图像的特征差异得到运动对象。同时,采用一种改进的背景维护方法以提高算法对环境光线变化的抗干扰能力。实验结果表明该算法具有快速、可靠的特点,可满足实时运动检测的需要。
Motion object detection is an important issue in the applications of computer vision. A new method to detect motion objects was presented, by comparing the texture diversity between the foreground image and the background image. The texture feature of an image block can be attained from its wavelet domain in different frequency. To adapt itself to the changing environment, an improved method to maintain the background was introduced. The experimental results show that the algorithm is fast and reliable, which meets the requirements of real-time motion object.
出处
《计算机应用》
CSCD
北大核心
2007年第1期237-239,共3页
journal of Computer Applications
基金
广东省重点攻关项目(2004B10101032)
关键词
运动对象
小波变换
纹理特征
motion object
wavelet transform
texture feature
