改进的背景减法与五帧差分法相结合的运动目标检测

Moving Object Detection of Improved Background Subtraction in Combination with Five Image Differential

为了能够更加准确、完整地检测出视频中的运动目标,该文基于改进的背景减法与五帧差分法,提出了一种高效、快捷的目标检测方法。该方法对视频中相邻五帧图像进行灰度变换、中值滤波和直方图均衡化等预处理,以降低噪声影响;利用五帧差分法得到第3帧中目标的大致范围的二值图像,同时将第3帧与系统预存背景帧差分,并对获得的结果实行边缘提取操作和二值化处理;将2种方法得到的图像进行逻辑或运算,即可得到运动目标的精确区域。经试验证明,该种方法获得的目标区域轮廓更加清晰,内容更加丰富。

In order to detect the moving object in the video more accurately and completely,an efficient and fast ob- ject detection method is proposed based on the improved background subtraction and the five frame difference method. Firstly,the five adjacent frames of the video are processed by gray level transfolTnation,median filtering and histogram equalization to reduce the influence of noise. And the binary image of the approximate range of the target in the third frame is obtained by the five frame difference method. At the same time,the third frame and back- ground frame are conducted frame difference,edge extraction operation and binarization processing are performed on the obtained result. The exact region of the moving object can be obtained by the logical OR operation. Experiments show that the method of obtaining the target area contour is more clear, more richer content.