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

Method for reducing smoke and fire monitoring calculation amount through attention mechanism and electronic equipment

A technology of attention and calculation, applied in the field of fire monitoring calculation and attention mechanism to reduce smoke, which can solve the problems of wasting computing power and having very little time.

Pending Publication Date: 2020-12-22
WENZHOU UNIVERSITY
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] At present, the resolutions of surveillance cameras are often 1080 (1920×1080), 3MP (2560×1440) and 5MP, but smoke and fire only occurred in local areas in the early days, and monitoring such a large format wastes computing power; on the other hand, normal There are many states, and only a few disasters

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
  • Method for reducing smoke and fire monitoring calculation amount through attention mechanism and electronic equipment
  • Method for reducing smoke and fire monitoring calculation amount through attention mechanism and electronic equipment
  • Method for reducing smoke and fire monitoring calculation amount through attention mechanism and electronic equipment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0050] This embodiment discloses spatial attention, in which the spatial attention mainly focuses on the area with rich effective information of the image, and adopts "smoke and fire detection convolution network based on static analysis features such as color, texture and edge", such as figure 1 As shown, first use the color feature channel to generate, and then use two branches to extract texture and edge features respectively, and then combine them into a comprehensive feature of 640 channels, and then maximum pooling and average pooling along the channel axis to obtain 2-channel features. Figure, and then use 7X7 convolution and sigmoid activation function to finally obtain a 1-dimensional weight feature map. The place where the value of the weight feature map is large is the place to pay attention. Among them, the smoke and fire detection convolutional network based on static analysis features such as color, texture, and edge is described below.

[0051] The color featur...

Embodiment 2

[0064] This embodiment discloses temporal attention. The process of occurrence and development of fireworks has temporal continuity. Therefore, spatial attention is embedded into the codec in the form of a recurrent neural network (RNN), such as Image 6 shown. Among them, it is first encoded by the encoder E, and then decoded by the decoder D. On the basis of the codec, an attention mechanism is added to form a time-series codec network with spatial attention, such as Figure 7 shown.

[0065] First measure the t-ith hidden state E of the encoder E t-i and previous decoder D state D t-1 to D t The contribution size of each encoder is continuously adjusted, so as to pay more attention to the parts similar to the smoke and fire features, while suppressing other useless information.

[0066] The following steps are involved when calculating attention:

[0067] S1 will E t-i (0≤i≤N) and each D t-1 Calculate the weight f(E in the perceptron way t-i ,D t-1 )f(E t-i ,D t...

Embodiment 3

[0073] This embodiment discloses an electronic device, including a processor and a memory storing execution instructions. When the processor executes the execution instructions stored in the memory, the processor executes an attention mechanism to reduce smoke and fire monitoring method of calculating quantities.

[0074] To sum up, the present invention introduces a spatial attention mechanism, and only cares about the high-probability areas of smoke and fire; introduces a temporal attention mechanism, only cares about the high-probability periods of smoke and fire. Therefore, only further strict detection of high-probability time and space can greatly reduce the consumption of computing power for full-scale and full-time monitoring.

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 relates to the technical field of fire detection, in particular to an attention method for reducing smoke and fire monitoring calculation amount and electronic equipment, and the methodcomprises the following steps: firstly obtaining a to-be-detected image, generating a color feature channel through a color feature network, and respectively extracting texture features and edge features through a texture feature network and an edge feature network; converting and merging the texture features and the edge features into comprehensive features by using a large convolution kernel, and then using a sigmoid activation function to generate a one-dimensional weight feature map; embedding time attention between encoders and decoders of a recurrent neural network RNN, measuring contribution of the ith hidden state Eti of an encoder E and the D state Dt1 of the decoder to Dt firstly, adjusting the weight of each historical encoder continuously, and subjecting only frames with the decoder Dt exceeding a threshold value to key detection. According to the invention, only high-possibility space-time further strict detection is carried out, and the consumption of computing power by full-width and full-time monitoring can be greatly reduced.

Description

technical field [0001] The invention relates to the technical field of fire detection, in particular to a method and an electronic device for reducing the calculation amount of smoke and fire monitoring by an attention mechanism. Background technique [0002] Deep learning avoids the dependence on manual labor in traditional methods to a large extent, and can automatically obtain high-level features that are difficult to obtain with traditional techniques. A reasonably designed nonlinear structure can also preserve the low-dimensional manifold in the smoke and can be used to generate finer ground truth, which brings a significant improvement to the smoke recognition task of various granularities. [0003] However, deep learning to detect smoke and fire currently has the serious shortcomings of "large deep learning model, heavy computing burden, and high application deployment cost", low cost performance, and low customer deployment willingness. Therefore, reducing the compu...

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): G06T7/90G06T7/40G06T7/136G06T7/13G06N3/08G06N3/04
CPCG06T7/13G06T7/40G06T7/90G06T7/136G06N3/084G06N3/048G06N3/045
Inventor 罗胜
Owner WENZHOU UNIVERSITY
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