基于路况视频的气象能见度检测方法与系统设计

Visibility detection method and system design based on traffic video

为充分利用高速公路沿线的监控摄像机,实现对低能见度天气的全程监测,在分析能见度值的计算模型的基础上,提出了利用图像对比度进行视频能见度检测的算法。算法采用一致对比度法的边缘特征提取出图像中能够反应能见度变化的图像特征参数,利用最小二乘法及逆变换得出对比度与距离的拟合函数,分析图像特征与距离之间的关系,确定能见度数值。算法较好地满足人眼视觉特性,有效地提高能见度检测的精度。通过在高速公路搭建系统进行实验与测试表明,该视频能见度检测系统实时性好,检测准确度高、误差小,且具有较好的鲁棒性,并能对全路段能见度情况进行实时统计与分析。

To make full use of traffic video cameras along highways and achieve low-visibility weather monitoring, based on the analysis of the calculation model for visibility value, a meteorological visibility detection method is proposed using image contrast. This method adopts the threshold value of the critical distance of the image visibility to calculate meteorological visibility. In order to extract out the image feature parameters that reflect the changes in visibility, the algorithm uses least squares method and inverse transform to draw the fitting function of contrast and distance, and analyzes the relation between the image characteristics and distance to determine the visibility values. The algorithm meets the visual characteristic of human eyes, and effectively improves the accuracy of visibility detection. The video visibility detection system has good real-time feature, high detection accuracy, small error and better robustness. The system can be used to carry out real-time statistics and analysis of the visibility of highways.