Subway vehicle number recognition method based on deep learning target detection
A target detection and deep learning technology, which is applied in neural learning methods, character recognition, character and pattern recognition, etc., can solve the problems of slow deep learning training speed, and the recognition accuracy needs to be improved. Recognize effects with high accuracy
Pending Publication Date: 2021-02-26
CHINA RAILWAY FIRST SURVEY & DESIGN INST GRP
View PDF5 Cites 2 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Most of the current methods use deep learning to directly find the car number on the image, and then use OCR technology to identify the car number. The car number is changing image information. The training speed of deep learning is slow, and the recognition accuracy needs to be improved.
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0024] The present invention will be described in detail below in combination with specific embodiments.
[0025] Such as figure 2 As shown, the front part of the subway vehicle is marked with the car number, and there must be a specific LOGO at the same time. Different cars have different car numbers, but the LOGO is generally the same, and the LOGO is a fixed target.
[0026] The invention relates to a subway car number recognition method based on deep learning object detection (Object Detection). Firstly, the LOGO is identified, and according to the positional relationship between the LOGO and the car number, the car number is found through the LOGO, and then the car number is identified and sent to Superior server. The process integrates deep learning target detection technology, machine vision technology and OCR technology (such as TESSERACT).
[0027] Deep learning (deep learning) is a branch of machine learning. It is an algorithm based on artificial neural network t...
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
Login to View More Abstract
The invention relates to a subway vehicle number recognition method based on deep learning target detection, and the method comprises the steps: obtaining vehicle head image information when a subwayvehicle passes, and recognizing a LOGO on the front side of a vehicle head based on a deep learning target detection technology; according to the relative position relation and the size proportion relation between the LOGO on the front face of the vehicle head and the vehicle number, calculating the position and the size of the vehicle number on the image based on the machine vision technology, finding the position of the vehicle number, performing text recognition through an OCR algorithm after the image is cut, and obtaining vehicle number data; and sending the vehicle number data to a superior server. The core innovation of the strategy is that the vehicle number is not directly identified, but the LOGO is firstly identified, the vehicle number is found through the LOGO according to theposition relationship between the LOGO and the vehicle number, and then the vehicle number is identified; and the LOGO is a fixed target, compared with the recognition accuracy of the vehicle number,the recognition accuracy is higher, fewer samples are needed during deep learning, and the training speed is higher.
Description
technical field [0001] The invention relates to the technical field of subway repair and maintenance, in particular to a subway number recognition method based on deep learning target detection. Background technique [0002] Many systems of subway maintenance units need to know the vehicle number information of incoming vehicles in real time in order to carry out work and record data, such as automatic car washing machines, wheel set flaw detection systems, pantograph online detection systems, etc. The level of these systems is often not up to the standard The requirements for direct access to the vehicle operation management signal system require accurate and rapid acquisition of vehicle number information from outside the signal system. [0003] If the subway vehicles are not retrofitted with the installation of RFID tags (usually not allowed), it is necessary to use vision technology to identify the vehicle number. Most of the current methods use deep learning to directl...
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
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06K9/34G06N3/08
CPCG06N3/08G06V20/40G06V30/153G06V30/10
Inventor 史时喜严飞周航博马森月丁子全侯小祥
Owner CHINA RAILWAY FIRST SURVEY & DESIGN INST GRP