Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Vehicle-mounted mobile terminal target detection method based on improved Yolov5

A vehicle-mounted mobile and target detection technology, which is applied in the direction of instruments, biological neural network models, character and pattern recognition, etc., can solve the problems of low detection accuracy, failure to guarantee recognition reliability, and detection speed that cannot meet real-time requirements. The effect of hardware resources, reduced inference cost, and inference speed remains the same

Pending Publication Date: 2021-08-13
HANGZHOU DIANZI UNIV
View PDF5 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Among them, Yolov5s has the least model parameters, but the detection accuracy is not high, and the reliability of recognition cannot be guaranteed in practical applications; and the detection speed of Yolov5l and Yolov5x, after the recognition accuracy rate is improved, cannot meet the real-time requirements

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
  • Vehicle-mounted mobile terminal target detection method based on improved Yolov5
  • Vehicle-mounted mobile terminal target detection method based on improved Yolov5
  • Vehicle-mounted mobile terminal target detection method based on improved Yolov5

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0017] The present invention is further explained below with reference to the accompanying drawings.

[0018] like figure 1 As shown, based on the improved YOLOV5, the vehicle mobile target detection method, specifically includes the following steps:

[0019] Step 1, add the full connect layer of the YOLOV5 model three output branches figure 2 The RFP module shown, the RFP module includes a multi-branch convolution layer and a branching cell, and three convolution layers in the multi-branch convolution layer parallel, sharing the same structure and weight, by different expansion ratios. Features provide different perceptual domains to change the sensation domain size of the output feature, allowing adaptives for each feature to have different perceptual domains, and improve the recognition accuracy of the multi-scale target. Among them, the multi-branch convolution layer is 3 parallel, the expansion rate of 1, 3, 5 is a expansion convolution of 3x3, respectively. The branch pool l...

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 vehicle-mounted mobile terminal target detection method based on improved Yolov5. According to the method, the Yov5 network is improved through an RFP (Reception Field Pyramid) module, and the RFP module is added behind a feature pyramid of a Neck part of the Yov5 network; and the improved network is trained by using the data set to obtain a weight model. And then the trained model is transplanted to a vehicle-mounted mobile terminal for real-time detection and identification. An RFP module is added behind a feature pyramid, so that a feature map originally output at a Neck layer by a Yolov5 network can adaptively have different sensory domains, the recognition precision of the Yolov5 network on a multi-scale target is improved, and meanwhile, the model is small in size and high in recognition speed. The problems that real-time detection cannot be carried out on a vehicle-mounted board due to the fact that a model is too large and the multi-scale target recognition accuracy is not high are solved.

Description

Technical field [0001] The present invention belongs to the field of target detection, and it is directed to a target detection method of a vehicle-mounted moving end, and is specifically involved in an in-vehicle moving end target detection method based on improved YOLOV5. Background technique [0002] The core of the unmanned driving system can be divided into three parts: perception, planning and control. Perception is to collect information from the vehicle's driving environment and extract related knowledge for later planning and control, is the basic link in the process of unmanned technology implementation. [0003] Traditional target detection methods are primarily based on characteristics. Features Learning Depending on the complexity of the model, feature selection and classification calculations can be divided into machine learning methods and depth learning algorithms. Traditional machine learning method, first perform regional selection, usually use a sliding window,...

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): G06K9/00G06K9/62G06N3/04
CPCG06V20/56G06N3/045G06F18/214
Inventor 高明裕王俊帆董哲康杨宇翔周洪涛王耀农赵志定
Owner HANGZHOU DIANZI UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com