Unlock instant, AI-driven research and patent intelligence for your innovation.

Attention-based convolutional neural network pointer type instrument image reading recognition method

A convolutional neural network and recognition method technology, which is applied in the field of attention-based convolutional neural network pointer meter image reading recognition, can solve the problem of large model parameters and space complexity, small size, and rough identification of meter readings. and other problems to achieve the effect of improving the type and distribution and simplifying the model volume

Active Publication Date: 2021-08-20
HUZHOU ELECTRIC POWER SUPPLY CO OF STATE GRID ZHEJIANG ELECTRIC POWER CO LTD
View PDF12 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to overcome the problems that the readings of the instrument recognition by the convolutional neural network in the prior art are difficult to directly depend on the information of the dial area, the reading method of the identification instrument is relatively crude, and the parameters and space complexity of the model are large, and provide a method for constructing the model volume Small, easy to deploy, high efficiency, and high prediction accuracy based on attention-based convolutional neural network pointer instrument image reading recognition method

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
  • Attention-based convolutional neural network pointer type instrument image reading recognition method
  • Attention-based convolutional neural network pointer type instrument image reading recognition method
  • Attention-based convolutional neural network pointer type instrument image reading recognition method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0045] Such as Figure 1-5 An attention-based convolutional neural network pointer meter image reading recognition method is shown, including the following steps:

[0046] Step 1: Make Faster R-CNN dataset. The pointer instrument image data set collected by the robot is divided into training set and test set according to the ratio of 4:1. For each picture in the training set, make two XML files: the first XML file records the position and label of the dial area. In order to obtain the correct type of instrument later, the label of the dial needs to indicate the specific model of the instrument, such as SF6 pressure gauge; The second XML file records the position and label of the pointer area, and the label is uniformly marked as a pointer. These two XML files will be used to train the Faster R-CNN network to detect and recognize the dial area and pointer area respectively.

[0047] Step 2: Build a Faster R-CNN network. Such as figure 2 The Faster R-CNN network is establi...

Embodiment 2

[0065] Embodiment 2: A kind of convolutional neural network substation meter image reading recognition method based on attention, select 1106 SF6 pressure gauge images of a certain substation as the experimental data set for automatic recognition of pointer meter readings, each image resolution is 256*256. The experimental data set is divided into 884 pictures as the training set and 222 pictures as the test set according to the ratio of about 4:1. The CPU of the test platform is Core i7-9700K, and the GPU is a single-core GEFORCE RTX 3090 Ti.

[0066] Define the accuracy rate of reading recognition as the ratio of the number of images whose absolute value of the relative error of reading recognition is within the scale range of one unit to the total number of images; define the deviation rate as the absolute value of the relative error of reading recognition within the range of one unit to two units of scale The ratio of the number of images to the total number of images; th...

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 attention-based convolutional neural network pointer type instrument image reading recognition method, which comprises the following steps: firstly, detecting a dial plate of an instrument image by using a Faster R-CNN, establishing a Faster-CNN data set of a pointer type instrument image, simultaneously obtaining the type of the dial plate, then training and testing by using an attention-based convolutional neural network model data set, after training is finished, performing hierarchical regression on the convolutional neural network based on the attention mechanism to obtain pointer readings. The attention module is introduced into the convolutional neural network, the extraction of instrument image features is enhanced by adopting the two-way heterogeneous convolutional neural network, and the type and distribution of the extracted features are improved by adding the convolutional attention module. The convolutional neural network adopts a hierarchical regression strategy, so that the model volume is greatly simplified. The convolution attention module improves the type and distribution of extracted features and improves the accuracy of meter reading.

Description

technical field [0001] The invention relates to the field of data recognition, in particular to an attention-based convolutional neural network pointer meter image reading recognition method. Background technique [0002] Due to the large amount of electromagnetic interference in the substation environment and the price advantage of pointer meters, there are still a large number of pointer meters in substations. The promotion of inspection robots in substations has greatly facilitated the acquisition of pointer instrument images. In order to further save labor costs and improve the automation and intelligence level of substations, it is necessary to improve the accuracy and efficiency of pointer instrument reading recognition in substation inspection images. [0003] Generally, the process of automatic recognition of pointer instrument readings is roughly divided into two steps: calibrating the dial and identifying the position of the pointer. Because the algorithm princip...

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): G06K9/20G06K9/34G06K9/46G06K9/62G06N3/04G06N3/08G06N20/10
CPCG06N3/08G06N20/10G06V10/22G06V10/267G06V10/44G06V2201/02G06N3/048G06N3/045G06F18/2411Y04S10/50
Inventor 管敏渊李凡归宇王涤徐凯杨斌戴则维杜鹏远赵崇娟王瑶黄宇宙闻俊义
Owner HUZHOU ELECTRIC POWER SUPPLY CO OF STATE GRID ZHEJIANG ELECTRIC POWER CO LTD