YOLOv5 neural network vehicle detection method added with attention mechanism

A neural network and vehicle detection technology, applied in the field of image detection and recognition, can solve the problems of no allocation, affecting image recognition, errors, etc., and achieve the effect of improving the recognition accuracy, the speed of model convergence, and the effect of improving the effect.

Pending Publication Date: 2022-02-25
YANGZHOU UNIV
View PDF0 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] Due to the complex road environment and changing background environment will affect the accuracy of image recognition, the image recognition detection method adopted in the prior art will be affected by the environment, and there is no targeted allocation of enough "attention" to the key parts of the training image. "power", sometimes misidentification or omission of identification occurs, and improving the accuracy and efficiency of identification has become an important problem that needs to be solved urgently

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • YOLOv5 neural network vehicle detection method added with attention mechanism
  • YOLOv5 neural network vehicle detection method added with attention mechanism
  • YOLOv5 neural network vehicle detection method added with attention mechanism

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0035] refer to Figure 1~6 , is an embodiment of the present invention, provides a kind of YOLOv5 neural network vehicle detection method that adds attention mechanism, comprises:

[0036] S1: Use the car image dataset and preprocess the dataset.

[0037] It should be noted that the UA-DETRAC public dataset is used, the image labeling tool Labelimg is used to label the image, and the COCO dataset format is saved as a txt format label file, and the dataset is divided into a training set according to a division ratio of 8:1. and validation set.

[0038] S2: Improve the YOLOv5 neural network and add an attention mechanism.

[0039] It should be noted that, if the channel attention module is added, the input feature map, that is, H×W×C, is passed through the global max pooling layer (global max pooling) and the global average based on H and W respectively. Pooling (global average pooling) to obtain two 1×1×C feature maps; send the feature maps to a two-layer neural network (ML...

Embodiment 2

[0058] This embodiment is another embodiment of the present invention. What this embodiment is different from the first embodiment is that it provides a verification test of the YOLOv5 neural network vehicle detection method that adds attention mechanism, which is the technology adopted in this method. The effect is verified and explained. In this embodiment, the traditional technical solution is used to conduct a comparative test with the method of the present invention, and the test results are compared by means of scientific demonstration to verify the real effect of the method.

[0059] In this example, the training CPU is Intel Core i5-6300HQ@2.30GHz, the GPU is GTX1080Ti, the deep learning framework used is Pytorch1.9.0, and the CUDA version is 11.2, and then follow the steps in Example 1 to input the image data into the unimproved YOLOv5 Train and test in the network, compare the test results of the two, and the results are shown in the following table:

[0060] Table 1...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a YOLOv5 neural network vehicle detection method added with an attention mechanism. The method comprises the steps: employing an vehicle image data set, and carrying out the preprocessing of the data set; the YOLOv5 neural network is improved, and an attention mechanism is added; inputting the marked data set into the improved YOLOv5 neural network according to a format meeting a network requirement for training and testing a result; and deploying the trained model to a mobile terminal to detect and identify a target vehicle. According to the method, the small target recognition effect of the model is improved, the model recognition precision and the model convergence speed are improved, and the real-time target recognition function of the mobile terminal is realized.

Description

technical field [0001] The invention relates to the technical field of image detection and recognition in artificial intelligence and traffic safety, in particular to a YOLOv5 neural network vehicle detection method adding an attention mechanism. Background technique [0002] In recent years, with the continuous advancement of science and technology, the application of artificial intelligence technology in many fields has achieved beneficial results, and the intelligent detection and recognition technology of images has also emerged. With the design and development of a series of neural network frameworks such as YOLO and Tensorflow, this This technology is also becoming more and more mature. If this image recognition technology is applied to the traffic safety system, when some road sections are congested, corresponding dredging measures can be quickly taken for the congested areas. When emergency vehicles such as ambulances or fire engines need to pass , the road can be cl...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06V10/774G06V10/764G06V10/82G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/241G06F18/214
Inventor 夏长权汪李超时壮壮朱颖徐思韵
Owner YANGZHOU UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap