运动图像目标检测指的是从序列图像中将变化的目标从背景中分离出来, 高斯混合模型可以对视频序列图像的前景和背景进行分类, 再利用背景减除实现运动目标的检测。提出一种基于改进高斯混合模型的优化背景建模方法, 该方法首先利用3×3模板对序列图像帧中的像素进行类似卷积的均值计算, 然后利用相邻均值的差提取均差因子自适应更新图像的均值。在此基础上, 设计了自适应学习率和学习速率, 利用改进高斯混合模型实现序列图像的背景建模。改进模型不仅能有效减少数据计算量, 同时可以降低在相似区域像素计算的时长, 大大加快背景建模速度。实验结果表明, 改进模型在目标检测、算法执行速率等性能指标上都有更好的表现, 能满足实时检测要求。

Moving image target detection refers to the separation of the changed target from the background from the sequence images. Gaussian mixture model can classify the foreground and background of the video sequence image, and then use background subtraction to achieve the detection of moving target. In this paper, an optimized background modeling method based on the improved Gaussian mixture model is proposed. Firstly, the average value of the pixels in the sequence image frame is calculated by using the template similar to convolution, and then the average value of the image is updated adaptively by using the difference of the adjacent average values. On this basis, the adaptive learning rate and learning rate are designed, and the improved Gaussian mixture model is used to realize the background modeling of sequence image. The improved model can not only effectively reduce the amount of data calculation, but also reduce the time of pixel calculation in similar areas, and greatly accelerate the speed of background modeling. Experimental results show that the improved model has better performance in target detection, algorithm execution rate and other performance indicators, and can meet the requirements of real-time detection.