Rider helmet attribute identification method

The invention discloses a rider helmet attribute recognition method, and belongs to the technical field of image processing. The method comprises the following steps: S1, a pedestrian attribute network being constructed, wherein the pedestrian attribute network comprises a backbone network and two sub-task networks, and the sub-task networks comprise a global network and a head segmentation network; S2, inputting the pedestrian picture into a pedestrian attribute network, obtaining a traffic mode attribute probability vector through a global network, and obtaining a helmet attribute probability vector through a head segmentation network; and S3, combining the traffic mode attributes and the helmet attributes to obtain pedestrian attributes. By means of the method, the traffic mode of the traveler and the helmet wearing situation can be efficiently and accurately recognized. 本发明公开了一种骑行人头盔属性识别方法，属于图像处理技术领域。该方法包括以下步骤：S1、构建行人属性网络，行人属性网络包括骨干网络和两个子任务网络，所述子任务网络包括全局网络和头部分割网络；S2、将行人图片输入行人属性网络，通过全局网络得到交通方式属性概率向量，通过头部分割