英语翻译The ability to reliably detect and track swimmers is ess

英语翻译
The ability to reliably detect and track swimmers is essential for successful swimmer motion analysis.However,a swimming pool contains many motion clutters including rapid fluctuation of water surface,water ripples/reflections,etc.This makes swimming pools different from most background scenes addressed by other video surveillance systems.To account for the unique non-stationary nature of swimming pools,we model the entire monitored pool area with a finite Gaussian mixture model so that rapid change of background pixels will be ignored and swimmers be detected
regardless of where they appear as long as they have low probability to be realized from the background model.A first order Kalman filter is then used to track each detected swimmer.The reader is directed to [1] for more details on swimmer detection and tracking.
For the sake of easy installation and maintenance,our system makes use of high-mounted overhead cameras to monitor swimmers.Every input frame is converted into a segmentation map,where foreground regions are motion blobs of the detected swimmers.Figure l shows an example of swimmer detection and tracking results.The first row shows the original frames from one test video clip,and in the second row pixels corresponding to the detected swimmers are highlighted in different colors to show the tracking of swimmers who moved in different directions and swimming styles.
先给这么多分 待会儿赚了在加
最好是能完整一点！

对一个成功的游泳选手进行动作分析,能够可靠地测定并跟踪是极为重要的.但游泳池池包含着许多的其他的无序动作如水表面的猛烈起伏,水的波纹以及反射回来的波等等.这使得游泳池和由其他的摄像监视系统所拍得的背景画面大不相同.为了解决游泳池的这唯一的不确定的因素,我们用有限的高斯混合模型来构建整个监视到的水池区域,背景像素的急剧变化就会淡化.这样无论游泳者身处何处,只要有从背景模型中认出他们的可能性,他们就会很容易地被识别出来.接下来一等的卡尔曼过滤器将被用来跟踪每一位被识别出的游泳者.关于游泳者的识别与跟踪,读者可以在指导下了解到更多的信息.
为了便于安装维护,我们的系统运用高置头顶的摄像头来监测游泳者.每个输入框架都被转换成分割图,图的前端区域即是被检测游泳选手的动作的模糊.图I显示的是一个游泳者的检测跟踪结果的实例.第一行显示了某一测试录像片段的原始框架.在第二行中,对应于被检测游泳者的像素则被不同颜色标示出,以对在水中向不同方向游动并不断变换方式的游泳者实施跟踪.