The invention discloses an infrared-based truck overload detection method, which specifically comprises the following steps of: installing a camera and an infrared thermal imager on the side of an expressway, and respectively detecting a truck and acquiring an infrared image; the camera detects a truck and recognizes the type of the truck through a Yolov5 deep learning algorithm based on an SA attention mechanism, and the thermal infrared imager collects feature images of an engine and tire parts of the truck; whether the truck is overloaded or not is judged through the ResNet algorithm, early warning is conducted on the overloaded truck, preliminary detection of highway truck overload is achieved under the non-contact condition, and potential safety hazards are reduced.
本发明公开了一种基于红外的货车超载检测方法,具体为:在高速公路路侧安装摄像头和红外热像仪,分别进行货车的检测和红外图像的获取;摄像头通过基于SA注意力机制的Yolov5深度学习算法检测货车和车型识别,红外热像仪采集货车发动机和轮胎部位特征图像;通过ResNet算法判断货车是否超载,对超载货车进行预警,在无接触的情况下实现高速公路货车超载的初步检测,减少安全隐患。
Infrared-based truck overload detection method
一种基于红外的货车超载检测方法
2024-06-07
Patent
Electronic Resource
Chinese
Truck overload detection system and method based on machine vision
European Patent Office | 2021
|Truck overload detection method based on radio frequency identification and target detection
European Patent Office | 2020
|Overload detection method and system for high-speed truck
European Patent Office | 2023
|Muck truck overload detection method and system based on intelligent network connection
European Patent Office | 2024
|