视频图像中运动目标检测算法的提高

摘要

The optimization of the moving target detection in complex environment which based on the Soldiers Shooting Confrontation System is proposed in this article. The algorithmis discussed from two aspects, the efficiency and the accuracy effected by the interference. To improve the efficiency of the algorithm, the large images from the video stream are first down-scaled,so that the location of the moving targets can be quickly determined.After that,the targets will be accurately analyzed in the original images according to their positions. In addition, the effect of the camera shaking and light changing can be eliminated by blending the current image with the background. Combining with background subtraction, background update can effectively improve the accuracy of the algorithm.%基于士兵对抗射击系统，研究复杂环境下运动目标检测算法的提高，从提高算法处理效率和降低外界干扰对算法准确性影响两方面进行讨论。对士兵对抗射击系统中大尺寸的视频流图像，先在下采样图像中判断运动目标的位置，再根据运动目标的位置，到原图中进行精确分析，从而提高算法的处理速度。通过背景融合方法更新图像背景，用不断更新的图像背景结合背景差分法获取运动目标，可以有效消除场景光线变化和相机抖动等因素对检测结果的影响，提高算法的准确性。