| Thick fog is a common catastrophic weather that has a very significant impact on social economy and people's production and life.Therefore,it is of great significance to accurately and timely forecast the fog weather for people's life.Texture is the salient feature of a weather situation map.Based on given weather situation map,it is a necessary premise on which online prediction model of fog pattern is carried out,to efficiently extract,represent and measure texture feature.Based on the analysis of the texture features of the e fog weather situation map,two real-time intelligent discriminant models for the fog weather are proposed:(1)The first one is the intelligent discriminant models for the fog weather based on ILBP(Improved Local Binary Pattern,ILBP),in which LBP operator in binary polynomial expression for the texture feature is improved as one in weight coefficient expression in order to carry out extracting the texture features along all kinds of different orientations.It is used to extract the texture features of the weather situation map and uses the Chi statistic function as a similarity measure function to discriminate the fog weather.(2)The second is the intelligent discriminant models for the fog weather based on image feature vector,in which texture features of the weather map are presented by calculating the feature vector after the block diagram of the fog weather situation map.The similarity of the two image feature vectors is calculated by using the similarity measure function based on the definition of the Euclidean distance to determine whether the two images are similar or not,thus to discriminate the fog weather.Taken the weather situation map dataset accumulated from 2010 to 2014 for Jiangsu region as benchmark dataset,whereas the weather situation map dataset accumulated from 2015 to 2017 for Jiangsu region as test dataset,furthermore the three indicators POD(Probability of Detection),FAR(False Alarm Rate)and CSI(Critical Success Index)as evaluation rulers,experimental results show the three indicators based on online prediction model of fog pattern are 0.884,0.15 and 0.76,and based on image feature vector of fog pattern are 0.871,0.169 and 0.74 respectively,these indicator values confirm that the proposed model has high accuracy and reliability. |
