Research on overload detection of trucks based on multi-sensor information fusion

This paper presents a novel approach for detecting truck overloads using a non-contact multi-sensor fusion method. The motivation behind this research stems from the limitations of existing highway truck overload detection systems, which suffer from low accuracy and challenging equipment maintenance. Initially, the collection of infrared temperature and sound data from the vehicle is facilitated through the utilization of infrared thermal camera and sound sensor devices that are positioned along the route. Second, the extracted infrared feature data are fused with the sound feature data for feature fusion, while principal component analysis (PCA) is utilized to reduce the data's dimensionality. Finally, the integration of the Flower Pollination method (FPA) and the Support Vector Machine (SVM) method is employed to develop a model for detecting truck overload. This model, known as EFPA-SVM, utilizes multi-feature fusion classification. The results of the tests indicate that the detection accuracy for identifying the true presence of 200 overloaded trucks is 82.3%. Additionally, it is feasible to achieve non-contact detection of lorry overload by combining temperature and auditory data. This research holds significance for the management of overloaded trucks.