基于RGB颜色空间的视频运动目标检测

为解决当前视频运动目标检测中检测精度不高以及视频颜色失真对运动目标检测的干扰问题,该文提出了一种改进的视频运动目标检测方法.比较了多种颜色空间下的运动目标检测算法,通过对视频的RGB(Red Green Blue)颜色空间建模,根据实际情况,对不同的颜色分量赋予不同的权值,提高了该颜色空间的真实性.同时,创新性地将神经网络与颜色空间结合,通过自组织映射,实现了对视频流数据中的运动目标检测.大量的实验结果表明,该方法对提高视频运动目标检测准确率有着显著的效果.

Moving Object Detection in Video Based on RGB Color Space

To solve the problem of low detection accuracy and color distortion in moving object detection,a moving object detection method based on RGB color space has been proposed.In this passage,kinds of moving object detection methods in different color spaces have been compared.Different weights been assigned to different color components in RGB color space to enhance the reality of the color space.Neural network has been combined with color space in this passage innovatively.By self-organizing map,this passage realizes the moving object detection in digital videos.A large numbers of experimental results showed that the proposed method can improve the accuracy of moving object detection.