The invention provides a traffic accident prediction method and device based on deep learning, and relates to the technical field of deep learning, and the method comprises the steps: obtaining a monitoring video corresponding to a target traffic scene; performing feature extraction on the monitoring video to obtain target image features; the target image features comprise RGB image global features, RGB image local features, depth image global features and depth image local features; obtaining association information between different features in the target image features by using a collaborative attention mechanism, and fusing the features in the target image features based on the association information to obtain a target fusion feature; and inputting the target fusion feature into a trained GRU network model, and outputting to obtain a traffic accident prediction probability corresponding to the target traffic scene. Through cascade combination of the collaborative attention layers, complementary perception fusion of multiple features is realized, and then a GRU module is utilized to predict whether traffic accidents occur in the future, so that the traffic accidents are reduced.
本发明提供一种基于深度学习的交通事故预测方法和装置,涉及深度学习技术领域,包括:获取目标交通场景对应的监控视频;对所述监控视频进行特征提取,得到目标图像特征;所述目标图像特征包括RGB图全局特征、RGB图局部特征、深度图全局特征和深度图局部特征;利用协同注意力机制获取所述目标图像特征中不同特征之间的关联信息,并基于所述关联信息将所述目标图像特征中的各个特征进行融合,得到目标融合特征;将所述目标融合特征输入至训练好的GRU网络模型,输出得到所述目标交通场景对应交通事故预测概率。通过协同注意力层的级联组合,以实现多特征的互补感知融合,进而利用GRU模块预测未来是否会发生交通事故,从而减少交通事故的发生。
Traffic accident prediction method and device based on deep learning
一种基于深度学习的交通事故预测方法和装置
2024-12-10
Patent
Electronic Resource
Chinese
Road traffic accident risk prediction deep learning algorithm
European Patent Office | 2022
|Traffic accident severity prediction based on interpretable deep learning model
Taylor & Francis Verlag | 2025
|Towards Deep Learning based Traffic Accident Analysis
IEEE | 2020
|European Patent Office | 2024
|