摘要
随着Android系统在智能手机、平板电脑等各种移动终端的广泛应用,Android成为具有广阔发展前景的开放性移动平台。由于Android智能手机和平板电脑的便携性,将运用OpenCV开发的行人检测系统移植到Android平台上,将改变传统的视频处理方式,使其更具灵活性、移动性,同时扩展了Android平台的应用领域。采用了基于高斯模型的背景差分法来提取前景,在跟踪部分运用了卡尔曼滤波来预测运动目标的搜索区域,根据几何校正后的前景面积来进行行人计数。系统在Android移动终端上测试,取得了较好的效果。
With the wide use of Android system in smart phones, tablet PCs and other mobile terminals, Android has become a mobile system which is an open source platform to the public and has broad development prospects. Because of the portability of Android smart phones and tablet PCs, transplanting pedestrian detection system which is developed with OpenCV to Android platform will change the traditional way of video processing, and will make the video processing become more flexible and more mobile. At the same time, transplanting video processing system to Android will expands the application fields of Android platform. This paper adopts background subtraction based on Gaussian model to extract foreground images,and then uses Kalman filter to predict the searching area of the moving objects in the tracking part. This paper uses geometric correction for all the foreground pixels,and then we predict the number of the pedestrian through the area of foreground images after geometric correction. The system is tested on Android terminals and has achieved good results.
出处
《电子测量技术》
2012年第12期99-101,116,共4页
Electronic Measurement Technology
