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

Automatic meter modeling method based on machine vision

A machine vision and automatic modeling technology, applied in neural learning methods, instruments, calculations, etc., can solve the problems of poor scalability, labor-intensive automatic recognition technology of pointer tables, and high errors

Inactive Publication Date: 2021-05-18
ZHEJIANG GUOZI ROBOT TECH
View PDF1 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0013] The purpose of the present invention is to overcome the disadvantages of the automatic recognition technology of pointer meters in the prior art, which is laborious and laborious, and has high error and poor expansibility, and provides an automatic meter modeling method based on machine vision

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
  • Automatic meter modeling method based on machine vision
  • Automatic meter modeling method based on machine vision
  • Automatic meter modeling method based on machine vision

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0067] Embodiment 1: a kind of meter automatic modeling method based on machine vision, such as figure 1 shown, including the following steps:

[0068] Step 1. Establish a meter master model library, which includes characteristic meter pictures, establish a meter header detection network using faster-rcnn and perform training, and establish a reid classification network and perform training;

[0069] Step 2: Input the image of the meter to be modeled, find the header image of the meter image to be modeled according to the meter header detection network, and then use the reid classification network to combine the image of the meter header of the meter image to be modeled and the characteristics of the master mold library Meter pictures for matching;

[0070] Step 3, after the matching is completed, arrange the characteristic meter pictures of the master mold database in order according to the confidence degree from high to low, and select the characteristic meter pictures of t...

Embodiment 2

[0099] Embodiment 2, a meter automatic modeling method based on machine vision, its principle and implementation method are different from Embodiment 1 in that:

[0100] The meter is a screw oil level gauge. In step 4, the key area indicated by the meter is the screw area and the central circle area. The matching includes the following steps ( Figure 5 ):

[0101] Step 4-5, select the candidate master mold with the highest confidence, and judge whether the number of screw types of the meter image of the meter to be modeled is consistent with the number of screw types of the candidate master mold with the highest confidence, and if they are consistent, skip Go to step 4-7, if not consistent, go to step 4-6;

[0102] Step 4-6, remove the candidate master model with the highest confidence, repeat steps 4-4 in the remaining candidate master models, if there is no candidate master model, it is judged that the matching fails;

[0103] Steps 4-7, respectively calculate the angle b...

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 an automatic meter modeling method based on machine vision, which overcomes the defects in the prior art. The method comprises the following steps of: 1, establishing a meter parent model library; 2, matching a header image of a to-be-modeled meter picture with a feature meter picture of the parent model library through a reid classification network; 3, selecting the feature meter pictures of the parent model library with the top several confidence coefficients as candidate parent models; 4, matching the header image of the to-be-modeled meter picture with the candidate parent models through a template of a meter indication key area, after matching is completed, determining final confidence coefficients by combining the confidence coefficients in the step 3, and determining that the candidate parent model with the highest final confidence coefficient is the optimal parent model matched with the to-be-modeled meter picture; and 5, calculating a homography matrix of the optimal parent model and the to-be-modeled meter picture, and converting the meter indication key area of the to-be-modeled meter picture according to the homography matrix to complete the modeling of a to-be-modeled meter.

Description

technical field [0001] The invention relates to the technical field of automatic modeling, in particular to a machine vision-based automatic meter modeling method. Background technique [0002] In some traditional industries, such as substations, chemical plants, oil refineries and other fields, various types of instruments are required to monitor data at any time to ensure the normal operation of each device. However, the existing meters require manual reading. Due to the large area and scattered equipment in most industrial places, it is difficult to find the abnormality of the meter value in a certain area in time when reading manually; in addition, manual work is not only costly but also efficient. Low also has great potential safety hazards. [0003] With the introduction of the smart grid concept, substation operation and operation and maintenance technologies are developing in the direction of unmanned and intelligent. Therefore, the current traditional method of re...

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): G06K9/20G06K9/32G06K9/46G06K9/62G06N3/04G06N3/08
CPCG06N3/04G06N3/08G06F18/22G06F18/2415
Inventor 林文益武诗洋李修亮余宗杰姚谦兰骏
Owner ZHEJIANG GUOZI ROBOT TECH
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