Traffic volume combination prediction model based on visual feature fusion
In the traffic volume combination prediction,there are problems of inaccurate traffic volume pre-diction and large prediction error,therefore,a traffic volume combination prediction model based on visual feature fusion is proposed based on big data.The time-based bilateral traffic flow series are calculated,and the time-series traffic volume data is analyzed.The identified color components of the visual image of traffic volume are grayed,the spatiotemporal association information is extracted,and the space-time sequence symbols are obtained.The temporal characteristic traffic volume sequence and the spatial characteristic traf-fic volume series are integrated,the visual features are fused by input,and the time-space series prediction weighting factor is adaptively adjusted to construct a traffic volume combination prediction model.It can be seen from the experiment results that the maximum error of weekdays'traffic volume prediction is 1％,and the prediction error of rest day is 0,which has an accurate prediction effect.
big datavisual feature fusiontraffic volumetime seriesspace sequence
万方数据
维普
