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

Flame detection method based on improved vibe algorithm and lightweight convolutional network

A flame detection, convolutional network technology, applied in the flame detection field of ViBe algorithm and lightweight convolutional network, can solve the problems of difficult flame area detection tasks, slow detection speed, difficult training, etc., to ensure speed and improve speed. , the effect of reducing the number of convolutional network layers

Active Publication Date: 2022-04-01
NANJING FORESTRY UNIV
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The existing deep convolutional network model for flame image segmentation has high complexity, difficult training and slow detection speed
Moreover, the public data sets used by these models do not contain flame objects, so it is difficult to apply to real-time flame area detection tasks

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
  • Flame detection method based on improved vibe algorithm and lightweight convolutional network
  • Flame detection method based on improved vibe algorithm and lightweight convolutional network
  • Flame detection method based on improved vibe algorithm and lightweight convolutional network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0024] Below in conjunction with accompanying drawing and specific embodiment the case is further described:

[0025] like figure 1 As shown, this method of flame area detection based on surveillance video mainly includes two steps, namely the improved ViBe foreground detection algorithm and the lightweight flame detection convolutional neural network based on the Squeeze-and-Excitation (SE) attention mechanism. Network construction. Specifically, it includes the following steps:

[0026] Step 1, utilize the improved ViBe prospect detection algorithm to carry out the prospect detection of flame, concrete steps are:

[0027] Step 1.1 selects the first frame of the video sequence, and utilizes the traditional ViBe algorithm to initialize the background model;

[0028] Step 1.2 defines the flickering degree of pixels in the video frame, as shown in equation (1):

[0029] BL t (x,y)=α×BL t-1 (x,y)+β×S or (x,y) (1)

[0030] Among them, BL t (x, y) indicates the degree of f...

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

A flame detection method based on an improved ViBe algorithm and a lightweight convolutional network, the steps of which are: using the improved ViBe foreground detection algorithm to detect the foreground of the flame; using the constructed lightweight convolutional neural network to further judge the detection of the ViBe foreground detection algorithm Whether the suspected flame area is a flame object, among which, a lightweight flame detection convolutional neural network based on the SE attention mechanism is constructed. The present invention adopts the ViBe foreground detection algorithm for the detection of suspected flame foreground targets, eliminates dynamic background noise points, and removes interference objects that are similar to flame colors but do not have flickering characteristics; compared with the existing deep convolution network for target detection, The invention improves the speed of flame detection; the SE attention mechanism is embedded in the lightweight flame detection convolutional network, which ensures the speed of flame detection while improving the accuracy of flame detection.

Description

technical field [0001] The invention belongs to the technical field of video detection, in particular to a flame detection method based on an improved ViBe algorithm and a lightweight convolutional network Background technique [0002] Video fire detection generally consists of two steps: moving foreground detection and flame recognition. However, in reality, there are many disturbances in complex environments, such as background disturbances (shaking leaves, fountains, etc.), light changes, shadow disturbances, etc., which bring challenges to the existing foreground detection algorithms. Once the foreground is detected, a flame recognition step is required. The fire recognition of traditional image processing extracts the geometry, texture, flicker frequency and other features of the flame area and designs a classifier to recognize the flame. Whether the feature selection is reasonable or not directly affects the accuracy of flame recognition. Applying deep learning techn...

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 Patents(China)
IPC IPC(8): G06V20/40G06V10/56G06V10/764G06V10/82G06V10/26G06K9/62G06N3/04G06N3/08
CPCG06N3/082G06V20/41G06V20/52G06V10/56G06N3/045G06F18/2415
Inventor 赵亚琴赵文轩卢鹏丁志鹏
Owner NANJING FORESTRY UNIV