机场能见度预测模型研究

Research on airport visibility prediction model

在雾霾天气下,基于图像处理的能见度检测方法仍然在不断研究中,对能见度估算值的精度依然具有提升空间。文章以大数据为基础,改进VGG卷积神经网络提取视频数据的特征并利用Adam进行算法优化,充分挖掘监控视频数据信息,以达到提高精度及降低设备成本的目的。相比于ResNet,这一方法充分利用了视频数据的时空信息,在预测过程中表现出较高的精度和准确性。这项研究对提升机场能见度预测的效果提供了借鉴。

In hazy weather,visibility detection methods based on image processing are still under continuous research,and the accuracy of visibility estimation is dependent on the accuracy of visibility estimation There is room for improvement.Based on big data,this paper improves VGG convolutional neural network to extract features of video data and uses Adam for algorithm optimization to fully mine surveillance video data information,so as to achieve the purpose of improving accuracy and reducing equipment cost.Compared with ResNet,this method makes full use of the spatio-temporal information of video data,and shows higher precision and accuracy in the prediction process.This study provides a reference for improving the effectiveness of airport visibility prediction.