基于改进均值滤波的浓雾中受困人员图像定位

针对传统定位方法准确率不高,耗费时间长的问题.提出了基于改进均值滤波的浓雾中受困人员实时定位算法,首先对浓雾环境进行视觉区域划分,对区域进行初始化操作,初始化操作后,对浓雾环境进行均值滤波处理,并对浓雾环境中的不同行像素跳变次数进行统计,浓雾图像中的跳变次数低于15的行是背景区域,高于15的行是受困人员像素区域.并引入空间位置参数,计算出浓雾中受困人员的位置坐标信息,采用先验边缘检测方法,借助边缘区域特征数,得到浓雾受困人员的准确空间位置.实验结果表明,改进方法比传统定位方法,定位准确率高,耗费时间短,实时性强,实际应用空间广阔.

A Study of Improved Median Filter of Real-Time Positioning for the People Trapped in Thick Fog

In view of the low accuracy and time-consuming nature of the traditional positioning method, a real-time positioning algorithm based on an improved median filter was proposed to locate the people trapped in thick fog. First, to visually zone the heavy fog environment, initialize regional operation, and then, to make average filtering processing of heavy fog environment, and to conduct a statistical investigation of the non-peer pixel jump frequency. Jump times lower than 15 lines of the heavy fog image is the background region and those over 15 lines are the personnel of the pixel area. With the introduction of spatial location parameters, the location coordinates of the people trapped in thick fog was calculated; with the prior edge detection method and the aid of edge number of regional characteristics, the accurate location of the people trapped in thick fog was acquired. The experimental results show that compared with the traditional positioning method, the improved method has a higher application.