视频图像中的运动目标检测

运动目标检测,是指从视频图像中将运动变化区域提取出来的检测技术,是图像处理技术的基础。在军事公安、交通管理、视频监控、医学检查等领域应用广泛。为了改进单独采用帧差法或背景减法进行运动目标检测时存在的不足,本文提出一种利用边缘信息的三帧差法与基于混合高斯模型的背景减法相结合的运动目标检测算法。该方法对视频图像中连续的三帧图像两两差分,对3个差分图像取均值,二值化,再经过形态学处理,并对中间帧进行Canny边缘提取,将二者进行"与"运算,即得到运动目标的边缘,用背景减法提取中间帧的前景,二值化,将其和目标的边缘进行"或"运算,经过形态学处理便可得到运动目标。实验结果表明,使用该方法目标检出率提高了9.7%~72.1%,误检率降低了0.090%~2.900%。这种二者相结合的方法相对于单一的检测算法能够有效、可靠地提取出运动目标。

Moving target detection of the video images

Moving target detection technology refers to extracting the change from the video images, and it is widely used in the field of military security, traffic management, video surveillance, medical examination and so on. In order to improve the shortcomings of three-frame-difference method and background subtraction, this paper presents a new moving object detection algorithm based on edge information of the three-frame-difference method combined with the background subtraction. Firstly, the current frame subtracts the previous frame and the next frame separately, the next frame subtracts the previous frame, their results are added together to get a gray-scale image.The result is divided by three. Secondly, the gray-scale image is translated into binary image after being judged by threshold. And then it will be dealt with a series of morphological processing.Thirdly, the Canny edge detection operator is introduced into the middle frame, then the two results "and" operation to get the edge of the moving object.Finally, the current frame image subtracts the background image to get another gray-scale image, in the same way, it is translated into binary image after being judged by threshold. Then the final two results "or" operation. The moving object is got by doing a series of morphological processing. The experimental results show that, target detection rate by using the method increased by 9.7%~72.1%, and the error detection rate is reduced by 0.090%~2.900%. The combination of the two compared with the single detection algorithm can effectively and reliably detect the moving targets.