The invention discloses a high-speed emergency lane occupation alarm method based on multi-mode large model area generation, and the method comprises the following steps: S10, building a multi-mode large model, building the multi-mode large model based on a Transform deep neural network, and carrying out the multi-mode large model; the multi-modal large model comprises a text information extraction module, a picture feature extraction module and a multi-modal feature fusion module; s20, collecting and processing picture data, collecting the picture data of the expressway, and manually labeling areas and describing picture information; s30, training a multi-modal large model; s40, emergency lane occupation warning is carried out; the method has the beneficial effects that the problems of high marking cost, low generalization performance, inflexibility and non-automation of emergency lane area generation in a high-speed scene are solved by using the multi-modal large model.
本发明公开了一种基于多模态大模型区域生成的高速应急车道占用告警方法,该方法包括以下步骤:S10、多模态大模型的搭建,基于Transformer深度神经网络搭建多模态大模型,所述多模态大模型包括文本信息提取模块、图片特征提取模块以及多模态特征融合模块;S20、图片数据的收集和处理,收集高速公路的图片数据,并通过人工标注区域和描述图片信息;S30、多模态大模型的训练;S40、应急车道占用告警;本发明的有益效果是:利用多模态大模型解决在高速场景下应急车道区域生成的标注成本高、泛化性能低、不灵活、不自动化的问题。
High-speed emergency lane occupation alarm method based on multi-mode large model region generation
基于多模态大模型区域生成的高速应急车道占用告警方法
2024-12-13
Patent
Electronic Resource
Chinese
Emergency lane occupation identification method and device
European Patent Office | 2021
|Emergency lane occupation behavior management system based on AI
European Patent Office | 2024
|European Patent Office | 2022
|Expressway emergency lane occupation detection system and implementation method thereof
European Patent Office | 2020
|