Flammable and explosive vehicle detection method and device in night environment, storage medium and equipment

By introducing FasterNet and Biformer mechanisms into the YOLOv12n network, the nighttime vehicle detection model was improved, solving the problem of low detection accuracy of flammable and explosive vehicles in nighttime environments, and achieving efficient vehicle identification and monitoring.

CN122313408APending Publication Date: 2026-06-30GUILIN UNIVERSITY OF TECHNOLOGY
0 Cites 0 Cited by

Patent Information

Application Number
CN202610413928.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-03-31
Publication Date
2026-06-30

AI Technical Summary

Technical Problem

At night or in low-light conditions, existing target detection algorithms struggle to effectively identify flammable and explosive vehicles, exhibiting low detection accuracy and insufficient feature extraction capabilities.

Method used

We introduce FasterNet feature extraction structure and Biformer attention mechanism into the YOLOv12n object detection network to improve the network structure and enhance feature extraction and fusion capabilities. The improved detection model is used to detect vehicle targets in nighttime images.

Benefits of technology

It improves the detection accuracy and stability of flammable and explosive vehicles in nighttime environments, effectively identifies dangerous transport vehicles such as oil tankers, and enhances road traffic safety monitoring capabilities.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses an intelligent detection method for flammable and explosive vehicles in a night environment, and relates to the technical field of target detection. According to the method, firstly, night road traffic images are collected, a vehicle detection data set is constructed, vehicle targets in the images are marked, and data preprocessing is carried out; then, a FasterNet feature extraction structure is introduced into a backbone network of the YOLOv12n target detection network to replace a part of convolution feature extraction modules, so that the network feature extraction capability is improved; meanwhile, a Biform attention mechanism is introduced between the output end of the feature fusion network and a detection head, and attention enhancement is carried out on the multi-scale fusion features. By training the improved model, flammable and explosive vehicles such as oil tank trucks and dangerous goods transport vehicles in a night traffic scene are detected, and the accuracy and stability of vehicle detection in a night environment are improved.
Need to check novelty before this filing date? Find Prior Art