基于数字摄像技术的能见度检测

Visibility detection algorithm based on data camera technology for road monitoring

为了更好地进行能见度的自动检测和全程检测,采用一种基于数字摄像技术的能见度检测方法:读入图像经过预处理设置"兴趣域",提取出对比度大于亮度对比感阈的人眼可分辨像素点,然后通过建立摄像机参数模型,计算出这些点到摄像机的距离,取其最大值即为能见度值。在自行研制的高速公路视频监控系统上,采集某段高速公路上的视频监控图像,进行能见度检测实验,得到923组的观测实验结果。实验结果表明,该算法与人眼观察一致,并对噪声有较好的鲁棒性。

A novel algorithm is presented to measure atmospheric visibility by video camera.Video cameras record the three-dimensional environment into two-dimensional data space by CCD.As a result,the depth information that exists in the three-dimensional environment,which is important for visibility detection,is lost.In order to overcome the limitation,interest domain is set.The eye distinguishable pixels are extracted through contrast calculation by the contrast threshold recommended by international commission on illumination.The distance from the pixels to the camera is calculated by the camera parameters through the model.The visibility is finally decided by the max.The contrast of the algorithm and the visibility sensor results based on variety scenes reveals that the two detectible methods are coherent,and the algorithm is effective and robust.