A vehicle type recognition method based on deep learning fusion model
By building a fusion model based on deep learning, the problem of insufficient generalization ability of traditional machine learning in vehicle type recognition is solved, and high-precision automatic recognition and information perception of vehicle types in highway scenarios are achieved.
CN116863412BActive Publication Date: 2025-09-30SOUTHEAST UNIV
0 Cites 0 Cited by
Patent Information
- Application Number
- CN202310664169.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-06-06
- Publication Date
- 2025-09-30
- Estimated Expiration
- 2043-06-06
AI Technical Summary
Technical Problem
Traditional machine learning methods have limited generalization capabilities in vehicle type recognition and are difficult to adapt to complex practical engineering needs. Existing technologies make it difficult to achieve accurate vehicle type recognition.
Method used
A deep learning-based fusion model is constructed, including CenterNet, VTR-NASNetLarge, VTR-VGG16, and VTR-MobileNetV2, to achieve automatic recognition of vehicle types through data enhancement and feature fusion.
Benefits of technology
It improves the accuracy and stability of vehicle type recognition, enhances the generalization ability of the model, and provides efficient perception of vehicle information in highway scenarios.
✦ Generated by Eureka AI based on patent content.
Abstract
The present invention discloses a vehicle type recognition method based on a deep learning fusion model, comprising: constructing a vehicle top view image dataset in a highway scene; constructing a vehicle type detection and recognition method CenterNet in a highway scene based on deep learning, and using the crop function to crop the vehicle target area; constructing a vehicle type recognition model VTR‑NASNetLarge based on deep learning, and obtaining a one-dimensional feature vector F N ; Build a deep learning-based vehicle type recognition model VTR-VGG16 and obtain a one-dimensional feature vector F V ; Build a deep learning-based vehicle type recognition model VTR‑MobileNetV2 and obtain a one-dimensional feature vector F M ; The eigenvector F N 、F V 、F M Parallel fusion is used to construct a deep learning-based vehicle type recognition fusion model, DFN-VTR, for toll vehicle type recognition in highway scenarios. This method captures the vehicle target area in the top view of the vehicle and constructs a deep learning fusion model. This allows for more accurate vehicle type recognition and provides technical support for vehicle information perception.
Need to check novelty before this filing date? Find Prior Art