Dangerous vehicle identification method based on cross attention mechanism dynamic knowledge propagation

A dangerous vehicle and identification method technology, which is applied in the field of dangerous vehicle identification based on the dynamic knowledge dissemination of the cross-attention mechanism, can solve the problem of high calculation volume, achieve the effect of improving detection performance, solving calculation volume and accuracy problems

Pending Publication Date: 2022-08-09
STATE GRID JIANGSU ELECTRIC POWER CO ELECTRIC POWER RES INST +2
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, even with a more efficient network, it can be computat

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
  • Dangerous vehicle identification method based on cross attention mechanism dynamic knowledge propagation
  • Dangerous vehicle identification method based on cross attention mechanism dynamic knowledge propagation
  • Dangerous vehicle identification method based on cross attention mechanism dynamic knowledge propagation

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0060] This embodiment introduces a method for identifying a dangerous vehicle, including:

[0061] Obtain the image frame of the dangerous vehicle to be identified;

[0062] Input the obtained image frames into the pre-trained dangerous vehicle recognition model;

[0063] According to the output of the dangerous vehicle identification model, determine whether there is a dangerous vehicle that makes a dangerous action in the image, and if so, output the dangerous vehicle information;

[0064] Among them, refer to figure 1 As shown, the dangerous vehicle recognition model includes a teacher network and a student network, the teacher network includes a first feature extraction network and a first temporal feature fusion network, and the student network includes a second feature extraction network, a second temporal feature fusion network, and a dynamic knowledge dissemination network. and prediction network; the first feature extraction network and the second feature extractio...

Embodiment 2

[0112] This embodiment introduces a dangerous vehicle identification device, including:

[0113] an image sampling module configured to acquire image frames of the dangerous vehicle to be identified;

[0114] an action classification module configured to input the acquired image frames into a pre-trained dangerous vehicle identification model;

[0115] an identification result output module, configured to determine whether there is a dangerous vehicle making dangerous actions in the image according to the output of the dangerous vehicle identification model, and if so, output dangerous vehicle information;

[0116] The dangerous vehicle identification model includes a teacher network and a student network, the teacher network includes a first feature extraction network and a first temporal feature fusion network, and the student network includes a second feature extraction network, a second temporal feature fusion network, and a dynamic knowledge dissemination network. networ...

Embodiment 3

[0125] This embodiment introduces a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method for identifying a dangerous vehicle as described in Embodiment 1 is implemented.

[0126] To sum up the above embodiments, the experiments show that the method of using the teacher model for reasoning in the present invention is effective for video recognition, and greatly improves the accuracy and robustness of dangerous vehicle recognition. And the complexity is low, the realization is simple, and the calculation speed is relatively fast.

[0127] As will be appreciated by those skilled in the art, the embodiments of the present application may be provided as a method, a system, or a computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the...

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 dangerous vehicle identification method and device and a storage medium, and the method comprises the steps: obtaining an image frame of a to-be-identified dangerous vehicle, and inputting a pre-trained dangerous vehicle identification model; according to the output of the dangerous vehicle identification model, determining whether a dangerous vehicle making a dangerous action exists in the image, and if yes, outputting dangerous vehicle information; wherein the dangerous vehicle identification model comprises a teacher network and a student network, the teacher network comprises a first feature extraction network and a first time feature fusion network, and the student network comprises a second feature extraction network, a second time feature fusion network, a dynamic knowledge propagation network and a prediction network; the dynamic knowledge propagation network adopts a cross attention mechanism to carry out feature cross fusion, then combines features obtained by the student network with cross participation features through residual connection, and further classifies the risk of vehicle actions in an image frame set. According to the invention, the dangerous vehicle can be identified, and the accuracy of the identification result can be improved.

Description

technical field [0001] The invention relates to the technical field of image detection, in particular to a dangerous vehicle identification method based on dynamic knowledge propagation of a cross-attention mechanism. Background technique [0002] With the vigorous construction of my country's power grid facilities, large-scale construction vehicles are widely used in power grid construction, which inevitably brings many problems in safety and management. In particular, it threatens the safe and stable operation of power transmission lines. Effective identification of dangerous vehicles has become the key to the successful extension of engineering vehicle identification technology to practical applications. [0003] The identification of dangerous vehicles is one of the key issues in the identification of large vehicles. At present, many mainstream vehicle detection methods are based on deep learning, and a variety of effective dangerous vehicle recognition methods have bee...

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
IPC IPC(8): G06V20/40G06K9/62G06V10/774G06V10/764G06V10/80
CPCG06F18/241G06F18/253G06F18/214Y04S10/50
Inventor 姚楠刘子全王真秦剑华朱雪琼薛海高超吴奇伟胡成博
Owner STATE GRID JIANGSU ELECTRIC POWER CO ELECTRIC POWER RES INST
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