CN116189100B - Gas hazard source detection and identification method and system based on spectral image - Google Patents
Gas hazard source detection and identification method and system based on spectral image Download PDFInfo
- Publication number
- CN116189100B CN116189100B CN202310466239.7A CN202310466239A CN116189100B CN 116189100 B CN116189100 B CN 116189100B CN 202310466239 A CN202310466239 A CN 202310466239A CN 116189100 B CN116189100 B CN 116189100B
- Authority
- CN
- China
- Prior art keywords
- degree direction
- feature
- coordinate system
- spectrum image
- directions
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/02—Alarms for ensuring the safety of persons
- G08B21/12—Alarms for ensuring the safety of persons responsive to undesired emission of substances, e.g. pollution alarms
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
- G06V10/12—Details of acquisition arrangements; Constructional details thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
- G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Emergency Management (AREA)
- Pathology (AREA)
- Immunology (AREA)
- Biochemistry (AREA)
- Analytical Chemistry (AREA)
- Chemical & Material Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Business, Economics & Management (AREA)
- Toxicology (AREA)
- Environmental & Geological Engineering (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a gas hazard source detection and identification method and system based on a spectral image, wherein the method comprises the following steps: acquiring a square area diagram of a target monitoring area; establishing a rectangular coordinate system by taking the position of the central point of the square area diagram as the origin of coordinates, and disposing a first spectrum image collector at the position of the origin of coordinates of the square area diagram; spectral image collectors are respectively deployed along 8 directions of a coordinate system; carrying out convolution operation feature extraction on multispectral image data acquired by the first spectral image acquisition unit; extracting features of n spectrum image collectors respectively deployed in 8 directions; calculating to obtain corresponding feature vectors in 8 directions; and fusing the feature vectors corresponding to the 8 directions, inputting the fused feature vectors into a pre-trained AlexNet recognition model, and obtaining a gas hazard source recognition result in the target monitoring area. Thereby improving the accuracy and efficiency of gas hazard source identification.
Description
Technical Field
The invention relates to the field of data processing, in particular to a gas hazard source detection and identification method and system based on a spectral image.
Background
In recent years, accidents caused by various dangerous gas explosions are more serious, the social requirement of informatization construction of the power-assisted Internet and safety production is met, the detection and identification of the gas dangerous sources are hot spot problems of current research aiming at chemical engineering leakage, and the accurate real-time detection and identification of the gas dangerous sources can provide effective help for reducing safety accidents and guaranteeing social safety. The prior art already has related technologies, for example, in CN107703555a (publication No. 20180216) by acquiring spectral image data of different bands of a preset protection area space; obtaining characteristic parameters of one or more dangerous sources in the preset protection area space according to the calculated difference or variation of the spectrum image data of different wave bands; and triggering an alarm after the one or more characteristic parameters are greater than or equal to a preset threshold value and last for a preset time. However, the detection method of the method is too single, the position arrangement of the spectrum image collector is irregular and can be found, unified standards are not formed, only a few gas dangerous sources can be identified, the accuracy and the efficiency of identification cannot be ensured, and the method has no obvious advantage in practical application.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide a gas hazard source detection and identification method and system based on a spectrum image, which are characterized in that a coordinate system is established, the data of a first spectrum image collector of a coordinate origin is taken as a reference, feature vectors are respectively extracted in the directions of 8 coordinate systems, the obtained 8 feature vectors corresponding to the directions are combined and fused, and then are input into a pre-trained AlexNet identification model, so that a gas hazard source identification result in a target monitoring area is obtained, and the detection can be carried out in real time, thereby improving the accuracy and efficiency of gas hazard source identification.
In order to achieve the aim of the invention, the invention adopts the following technical scheme:
in a first aspect, the present invention provides a method for detecting and identifying a gas hazard source based on a spectral image, the method comprising:
step 1, carrying out standardization processing on a target monitoring area to obtain a square area diagram of the target monitoring area;
step 2, taking the center point position of the square area diagram as a coordinate origin, taking the horizontal direction as an X axis and the vertical direction as a Y axis, and completing the establishment of a rectangular coordinate system;
step 3, disposing a first spectrum image collector at the origin position of coordinates of the square region map;
step 4, respectively disposing a spectrum image collector along the 0 degree direction, 45 degree direction, 90 degree direction, 135 degree direction, 180 degree direction, 225 degree direction, 270 degree direction and 315 degree direction of the coordinate system; the method specifically comprises the following steps:
respectively disposing n spectrum image collectors along the 0 degree direction, 45 degree direction, 90 degree direction, 135 degree direction, 180 degree direction, 225 degree direction, 270 degree direction and 315 degree direction of the coordinate system in the range of the square region graph according to the principle of fixed distance average distribution;
step 5, carrying out convolution operation feature extraction on the multispectral image data acquired by the first spectrum image acquisition device to obtain a feature map of the multispectral image, wherein the feature map comprises first features;
step 6, respectively extracting features of n spectrum image collectors respectively deployed in 8 directions, specifically including: respectively carrying out convolution operation feature extraction on the multispectral image data acquired by the n multispectral image collectors in a certain direction to obtain feature graphs of the n multispectral images corresponding to the certain direction, wherein the feature graphs comprise n corresponding second features;
step 7, performing a difference processing operation on the n corresponding second features and the first features to obtain a feature vector p= (x) corresponding to the certain direction 1 ,x 2 ,...,x n )；
Step 8, repeating the step 6 and the step 7 to obtain the corresponding feature vectors in 8 directions respectively;
step 9, merging and fusing the obtained 8 directional corresponding feature vectors to obtain a fused 8×n feature matrix Q, whereinWherein->Representing a first eigenvector calculated along the 0 deg. direction of said coordinate system,/i>Representing an nth feature vector calculated along a 0 ° direction of the coordinate system; />Representing a first eigenvector calculated along the 315 ° direction of said coordinate system,/->Representing an nth feature vector calculated along a 315 ° direction of the coordinate system;
and step 10, inputting the fused characteristic matrix Q into a pre-trained AlexNet recognition model to obtain a gas hazard source recognition result in the target monitoring area.
Further, the method further comprises the following steps: step 1, performing standardization processing on a target monitoring area to obtain a square area diagram of the target monitoring area, wherein the method specifically comprises the following steps:
and carrying out square frame identification processing according to the edge point information of the target monitoring area so as to obtain a square area diagram of the target monitoring area.
Further, the method further comprises the following steps: according to the length of the suitable monitoring distance of the currently selected spectrum image collector and the distance from the position of the origin of coordinates to the edges of the square area graph in 8 directions, respectively deploying n spectrum image collectors according to the principle of fixed distance average distribution.
Further, the method further comprises the following steps: the first feature and the second feature are specific features corresponding to specific identification gas in the multispectral image, and specifically comprise radiation brightness data, radiation temperature data and temperature difference data of specific wave bands.
Further, the method further comprises the following steps: the gas hazard source identification result comprises gas hazard source types, positions and relevant concentration information.
Further, the method further comprises the following steps: and carrying out grading alarm based on the gas hazard source identification result, and sending gas hazard source emergency treatment information to terminal equipment.
Further, the method further comprises the following steps: the convolution operation feature extraction in the step 5 and the step 6 specifically includes: and inputting the preprocessed multispectral image data with uniform size into a convolutional neural network for feature extraction, and obtaining a feature map of the multispectral image.
Further, the method further comprises the following steps: the principle of the fixed distance average distribution specifically comprises the following steps:
and respectively arranging n spectrum image collectors at equal intervals along the 8 directions from the origin position of coordinates to line segments at the edges of the square region graph.
Further, the method further comprises the following steps: the AlexNet recognition model is specifically obtained by performing iterative training by using a MapReduce framework.
In a second aspect, the present invention further provides a gas hazard source detection and identification system based on a spectral image, the system comprising:
the standardized processing module is used for carrying out standardized processing on the target monitoring area to obtain a square area diagram of the target monitoring area;
the rectangular coordinate system establishment module is used for completing rectangular coordinate system establishment by taking the center point position of the square regional graph as a coordinate origin, taking the horizontal direction as an X axis and taking the vertical direction as a Y axis;
the first deployment module is used for deploying a first spectrum image collector at the position of the origin of coordinates of the square region diagram;
the second deployment module is used for deploying the spectrum image collector along the 0 degree direction, the 45 degree direction, the 90 degree direction, the 135 degree direction, the 180 degree direction, the 225 degree direction, the 270 degree direction and the 315 degree direction of the coordinate system respectively; the method specifically comprises the following steps:
respectively disposing n spectrum image collectors along the 0 degree direction, 45 degree direction, 90 degree direction, 135 degree direction, 180 degree direction, 225 degree direction, 270 degree direction and 315 degree direction of the coordinate system in the range of the square region graph according to the principle of fixed distance average distribution;
the first feature extraction module is used for carrying out convolution operation feature extraction on the multispectral image data acquired by the first spectrum image acquisition device to obtain a feature map of the multispectral image, wherein the feature map comprises first features;
the second feature extraction module is used for respectively extracting features of n spectrum image collectors respectively deployed in 8 directions, and specifically comprises the following steps: respectively carrying out convolution operation feature extraction on the multispectral image data acquired by the n multispectral image collectors in a certain direction to obtain feature graphs of the n multispectral images corresponding to the certain direction, wherein the feature graphs comprise n corresponding second features;
a first feature vector calculation module for performing difference processing operation on the n corresponding second features and the first features to obtain feature vectors p= (x) corresponding to the certain direction 1 ,x 2 ,...,x n )；
The second feature vector calculation module is used for repeating the steps of the second feature extraction module and the first feature vector calculation module to respectively obtain feature vectors corresponding to 8 directions;
the fusion module is used for merging and fusing the obtained 8 feature vectors corresponding to the directions to obtain a fused 8 multiplied by n feature matrix Q, whereinWherein->Representing a first eigenvector calculated along the 0 deg. direction of said coordinate system,/i>Representing an nth feature vector calculated along a 0 ° direction of the coordinate system; />Representing a first eigenvector calculated along the 315 ° direction of said coordinate system,/->Representing an nth feature vector calculated along a 315 ° direction of the coordinate system;
and the recognition module is used for inputting the fused characteristic matrix Q into a pre-trained AlexNet recognition model to obtain a gas hazard source recognition result in the target monitoring area.
The beneficial effects are that:
1. in the invention, through step 1, a target monitoring area is subjected to standardization processing to obtain a square area diagram of the target monitoring area; step 2, taking the center point position of the square area diagram as a coordinate origin, taking the horizontal direction as an X axis and the vertical direction as a Y axis, and completing the establishment of a rectangular coordinate system; step 3, disposing a first spectrum image collector at the origin position of coordinates of the square region map; step 4, respectively disposing a spectrum image collector along the 0 degree direction, 45 degree direction, 90 degree direction, 135 degree direction, 180 degree direction, 225 degree direction, 270 degree direction and 315 degree direction of the coordinate system; step 5, carrying out convolution operation feature extraction on the multispectral image data acquired by the first spectrum image acquisition device to obtain a feature map of the multispectral image, wherein the feature map comprises first features; step 6, respectively extracting features of n spectrum image collectors respectively deployed in 8 directions, specifically including: respectively carrying out convolution operation feature extraction on the multispectral image data acquired by the n multispectral image collectors in a certain direction to obtain feature graphs of the n multispectral images corresponding to the certain direction, wherein the feature graphs comprise n corresponding second features; step 7, performing difference processing operation on the n corresponding second features and the first features respectively to obtain feature vectors corresponding to a certain direction; step 8, repeating the step 6 and the step 7 to obtain the corresponding feature vectors in 8 directions respectively; step 9, merging and fusing the obtained feature vectors corresponding to the 8 directions to obtain a fused 8×n feature matrix Q; and step 10, inputting the fused characteristic matrix Q into a pre-trained AlexNet recognition model to obtain a gas hazard source recognition result in the target monitoring area. According to the method, the coordinate system is established, the data of the first spectrum image collector of the coordinate origin is used as a reference, the feature vectors are respectively extracted in the directions of 8 coordinate systems, the obtained feature vectors corresponding to the 8 directions are combined and fused and then input into the pre-trained AlexNet recognition model, the gas dangerous source recognition result in the target monitoring area is obtained, and the detection can be carried out in real time, so that the accuracy and the efficiency of gas dangerous source recognition are improved. The method provides high reference value for the position arrangement of the spectrum image collectors in the protection area to form unified specification, and the applicable spectrum image collector model can be selected for various gases, so that the method can monitor various gases and has high applicability; meanwhile, the gas dangerous source emergency treatment information is sent to the terminal equipment through a grading early warning mechanism, and the method has important significance for fire rescue and emergency.
Drawings
Fig. 1 is a schematic diagram of steps of a gas hazard detection and identification method based on a spectral image.
Fig. 2 is a schematic diagram of an established rectangular coordinate system.
Detailed Description
The present application is described in further detail below with reference to the drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be noted that, for convenience of description, only the portions related to the invention are shown in the drawings.
As shown in fig. 1-2, the present embodiment provides a gas hazard source detection and identification method based on a spectral image, which includes:
step 1, carrying out standardization processing on a target monitoring area to obtain a square area diagram of the target monitoring area;
step 2, taking the center point position of the square area diagram as a coordinate origin, taking the horizontal direction as an X axis and the vertical direction as a Y axis, and completing the establishment of a rectangular coordinate system;
step 3, disposing a first spectrum image collector at the origin position of coordinates of the square region map;
step 4, respectively disposing a spectrum image collector along the 0 degree direction, 45 degree direction, 90 degree direction, 135 degree direction, 180 degree direction, 225 degree direction, 270 degree direction and 315 degree direction of the coordinate system; the method specifically comprises the following steps:
respectively disposing n spectrum image collectors along the 0 degree direction, 45 degree direction, 90 degree direction, 135 degree direction, 180 degree direction, 225 degree direction, 270 degree direction and 315 degree direction of the coordinate system in the range of the square region graph according to the principle of fixed distance average distribution;
step 5, carrying out convolution operation feature extraction on the multispectral image data acquired by the first spectrum image acquisition device to obtain a feature map of the multispectral image, wherein the feature map comprises first features;
step 6, respectively extracting features of n spectrum image collectors respectively deployed in 8 directions, specifically including: respectively carrying out convolution operation feature extraction on the multispectral image data acquired by the n multispectral image collectors in a certain direction to obtain feature graphs of the n multispectral images corresponding to the certain direction, wherein the feature graphs comprise n corresponding second features;
step 7, performing a difference processing operation on the n corresponding second features and the first features to obtain a feature vector p= (x) corresponding to the certain direction 1 ,x 2 ,...,x n )；
Step 8, repeating the step 6 and the step 7 to obtain the corresponding feature vectors in 8 directions respectively;
step 9, merging and fusing the obtained 8 directional corresponding feature vectors to obtain a fused 8×n feature matrix Q, whereinWherein->Representing a first eigenvector calculated along the 0 deg. direction of said coordinate system,/i>Representing an nth feature vector calculated along a 0 ° direction of the coordinate system; />Representing a first eigenvector calculated along the 315 ° direction of said coordinate system,/->Representing an nth feature vector calculated along a 315 ° direction of the coordinate system;
and step 10, inputting the fused characteristic matrix Q into a pre-trained AlexNet recognition model to obtain a gas hazard source recognition result in the target monitoring area.
In an alternative embodiment, the method further comprises: step 1, performing standardization processing on a target monitoring area to obtain a square area diagram of the target monitoring area, wherein the method specifically comprises the following steps:
and carrying out square frame identification processing according to the edge point information of the target monitoring area so as to obtain a square area diagram of the target monitoring area.
Specifically, square frame identification processing is performed on the principle that all edge points of the target monitoring area are just covered.
In an alternative embodiment, the method further comprises: according to the length of the suitable monitoring distance of the currently selected spectrum image collector and the distance from the position of the origin of coordinates to the edges of the square area graph in 8 directions, respectively deploying n spectrum image collectors according to the principle of fixed distance average distribution.
Specifically, the monitoring distances of the corresponding dangerous gases of the spectrum image collectors of different models are also different, so that in combination with the applicable monitoring distance length of the currently selected spectrum image collector, n spectrum image collectors are respectively deployed in a square area diagram along the 0-degree direction, 45-degree direction, 90-degree direction, 135-degree direction, 180-degree direction, 225-degree direction, 270-degree direction and 315-degree direction of the coordinate system, wherein n is more than or equal to 1.
In an alternative embodiment, the method further comprises: the first feature and the second feature are specific features corresponding to specific identification gases in the multispectral image;
specifically, the specific features are all relevant target features which are beneficial to target gas identification, and can be specific wave band radiation brightness data, radiation temperature data, temperature difference data and the like.
In an alternative embodiment, the method further comprises: the gas hazard source identification result comprises gas hazard source types, positions and relevant concentration information.
In an alternative embodiment, the method further comprises: and carrying out grading alarm based on the gas hazard source identification result, and sending gas hazard source emergency treatment information to terminal equipment.
Specifically, for each type of gas, a corresponding first-aid processing method is preset and stored in a database, and data in the database is directly called to alarm in critical time.
Specifically, the dangerous gas aimed at by the invention comprises: methane, propane, ethane, butane, benzene, ethylbenzene, toluene, gasoline, xylene, dimethylamine, methyl t-butyl ether, ethylene glycol butyl ether, diethyl ether, acetic acid, methanol, methyl ether, ammonia, sulfur dioxide, carbon dioxide, phenol, aniline, dichlorobenzene, styrene, nitrobenzene, peracetic acid, ammonium nitrate, boron trifluoride, allylamine, and the like. Meanwhile, the model selection of the spectrum image collector can be specifically selected according to specific dangerous gases to be monitored.
In an alternative embodiment, the method further comprises: the convolution operation feature extraction in the step 5 and the step 6 specifically includes: and inputting the preprocessed multispectral image data with uniform size into a convolutional neural network for feature extraction, and obtaining a feature map of the multispectral image.
In an alternative embodiment, the method further comprises: the principle of the fixed distance average distribution specifically comprises the following steps:
and respectively arranging n spectrum image collectors at equal intervals along the 8 directions from the origin position of coordinates to line segments at the edges of the square region graph.
Specifically, for example, according to the 0-degree direction of a coordinate system, n spectrum image collectors are arranged at equal intervals between the coordinate origin position and line segments of the rightmost edge of the square area graph according to the applicable monitoring distance length of the currently selected spectrum image collector and the distance from the coordinate origin position to the edge of the square area graph in the 0-degree direction.
According to the 45-degree direction of the coordinate system, according to the length of the applicable monitoring distance of the currently selected spectrum image collector and the distance from the position of the origin of coordinates to the edge of the square area diagram in the 45-degree direction, n spectrum image collectors are arranged between the position of the origin of coordinates to the line segments of the upper right corner of the edge of the square area diagram at equal intervals.
And referring to the method in other directions, disposing n spectrum image collectors at equal intervals.
In an alternative embodiment, the method further comprises: the AlexNet recognition model is specifically obtained by performing iterative training by using a MapReduce framework.
Specifically, an AlexNet neural calculation model is taken as a basic model, and a corresponding model conforming to a MapReduce calculation rule is constructed by combining the task type and the data quantity to be identified.
Based on the same inventive concept, the present embodiment provides a gas hazard source detection and identification system based on a spectral image, the system comprising:
the standardized processing module is used for carrying out standardized processing on the target monitoring area to obtain a square area diagram of the target monitoring area;
the rectangular coordinate system establishment module is used for completing rectangular coordinate system establishment by taking the center point position of the square regional graph as a coordinate origin, taking the horizontal direction as an X axis and taking the vertical direction as a Y axis;
the first deployment module is used for deploying a first spectrum image collector at the position of the origin of coordinates of the square region diagram;
the second deployment module is used for deploying the spectrum image collector along the 0 degree direction, the 45 degree direction, the 90 degree direction, the 135 degree direction, the 180 degree direction, the 225 degree direction, the 270 degree direction and the 315 degree direction of the coordinate system respectively; the method specifically comprises the following steps:
respectively disposing n spectrum image collectors along the 0 degree direction, 45 degree direction, 90 degree direction, 135 degree direction, 180 degree direction, 225 degree direction, 270 degree direction and 315 degree direction of the coordinate system in the range of the square region graph according to the principle of fixed distance average distribution;
the first feature extraction module is used for carrying out convolution operation feature extraction on the multispectral image data acquired by the first spectrum image acquisition device to obtain a feature map of the multispectral image, wherein the feature map comprises first features;
the second feature extraction module is used for respectively extracting features of n spectrum image collectors respectively deployed in 8 directions, and specifically comprises the following steps: respectively carrying out convolution operation feature extraction on the multispectral image data acquired by the n multispectral image collectors in a certain direction to obtain feature graphs of the n multispectral images corresponding to the certain direction, wherein the feature graphs comprise n corresponding second features;
a first feature vector calculation module for performing difference processing operation on the n corresponding second features and the first features to obtain feature vectors p= (x) corresponding to the certain direction 1 ,x 2 ,...,x n )；
The second feature vector calculation module is used for repeating the steps of the second feature extraction module and the first feature vector calculation module to respectively obtain feature vectors corresponding to 8 directions;
the fusion module is used for merging and fusing the obtained 8 feature vectors corresponding to the directions to obtain a fused 8 multiplied by n feature matrix Q, whereinWherein->Representing a first eigenvector calculated along the 0 deg. direction of said coordinate system,/i>Representing an nth feature vector calculated along a 0 ° direction of the coordinate system; />Representing a first eigenvector calculated along the 315 ° direction of said coordinate system,/->Representing an nth feature vector calculated along a 315 ° direction of the coordinate system;
and the recognition module is used for inputting the fused characteristic matrix Q into a pre-trained AlexNet recognition model to obtain a gas hazard source recognition result in the target monitoring area.
According to the invention, the coordinate system is established, the data of the first spectrum image collector of the coordinate origin is taken as a reference, the feature vectors are respectively extracted in the directions of 8 coordinate systems, the obtained feature vectors corresponding to the 8 directions are combined and fused and then input into the pre-trained AlexNet recognition model, so that the gas hazard source recognition result in the target monitoring area is obtained, the detection can be carried out in real time, the accuracy and the efficiency of gas hazard source recognition are improved, and the method is suitable for the detection of various gases; meanwhile, the gas dangerous source emergency treatment information is sent to the terminal equipment through a grading early warning mechanism, and the method has important significance for fire rescue and emergency.
Claims (10)
1. The gas hazard source detection and identification method based on the spectral image is characterized by comprising the following steps of:
step 1, carrying out standardization processing on a target monitoring area to obtain a square area diagram of the target monitoring area;
step 2, taking the center point position of the square area diagram as a coordinate origin, taking the horizontal direction as an X axis and the vertical direction as a Y axis, and completing the establishment of a rectangular coordinate system;
step 3, disposing a first spectrum image collector at the origin position of coordinates of the square region map;
step 4, respectively disposing a spectrum image collector along the 0 degree direction, 45 degree direction, 90 degree direction, 135 degree direction, 180 degree direction, 225 degree direction, 270 degree direction and 315 degree direction of the coordinate system; the method specifically comprises the following steps:
respectively disposing n spectrum image collectors along the 0 degree direction, 45 degree direction, 90 degree direction, 135 degree direction, 180 degree direction, 225 degree direction, 270 degree direction and 315 degree direction of the coordinate system in the range of the square region graph according to the principle of fixed distance average distribution;
step 5, carrying out convolution operation feature extraction on the multispectral image data acquired by the first spectrum image acquisition device to obtain a feature map of the multispectral image, wherein the feature map comprises first features;
step 6, respectively extracting features of n spectrum image collectors respectively deployed in 8 directions, specifically including: respectively carrying out convolution operation feature extraction on the multispectral image data acquired by the n multispectral image collectors to obtain feature graphs of n multispectral images corresponding to the direction, wherein the feature graphs comprise n corresponding second features;
step 7, performing a difference processing operation on the n corresponding second features and the first features to obtain a feature vector p= (x) corresponding to the direction 1 ,x 2 ,...,x n )；
Step 8, repeating the step 6 and the step 7 to obtain the corresponding feature vectors in 8 directions respectively;
step 9, merging and fusing the obtained 8 directional corresponding feature vectors to obtain a fused 8×n feature matrix Q, whereinWherein->Representing a first eigenvector calculated along the 0 deg. direction of said coordinate system,/i>Representing an nth feature vector calculated along a 0 ° direction of the coordinate system; />Representing a first eigenvector calculated along the 315 ° direction of said coordinate system,/->Representing an nth feature vector calculated along a 315 ° direction of the coordinate system;
and step 10, inputting the fused characteristic matrix Q into a pre-trained AlexNet recognition model to obtain a gas hazard source recognition result in the target monitoring area.
2. The method according to claim 1, wherein the step 1 of normalizing the target monitoring area to obtain a square area map of the target monitoring area specifically includes:
and carrying out square frame identification processing according to the edge point information of the target monitoring area so as to obtain a square area diagram of the target monitoring area.
3. The method according to claim 1, wherein the first and second features are specific features of the multispectral image corresponding to a specific identification gas, in particular radiation brightness data, radiation temperature data and temperature difference data of a specific wavelength band.
4. The method as recited in claim 1, further comprising:
according to the length of the suitable monitoring distance of the currently selected spectrum image collector and the distance from the position of the origin of coordinates to the edges of the square area graph in 8 directions, respectively deploying n spectrum image collectors according to the principle of fixed distance average distribution.
5. The method of claim 1, wherein the gas hazard identification results include gas hazard type, location and associated concentration information.
6. The method of claim 5, wherein the step of alarming is performed based on the gas hazard identification result, and the gas hazard emergency treatment information is transmitted to the terminal device.
7. The method according to claim 1, wherein the convolution operation feature extraction in the step 5 and the step 6 specifically comprises: and inputting the preprocessed multispectral image data with uniform size into a convolutional neural network for feature extraction, and obtaining a feature map of the multispectral image.
8. The method according to claim 4, wherein the principle of the fixed distance average distribution specifically comprises:
and respectively arranging n spectrum image collectors at equal intervals along the 8 directions from the origin position of coordinates to line segments at the edges of the square region graph.
9. The method according to claim 1, wherein the AlexNet recognition model is specifically obtained by iterative training using a MapReduce framework.
10. A gas hazard source detection and identification system based on spectral images, the system comprising:
the standardized processing module is used for carrying out standardized processing on the target monitoring area to obtain a square area diagram of the target monitoring area;
the rectangular coordinate system establishment module is used for completing rectangular coordinate system establishment by taking the center point position of the square regional graph as a coordinate origin, taking the horizontal direction as an X axis and taking the vertical direction as a Y axis;
the first deployment module is used for deploying a first spectrum image collector at the position of the origin of coordinates of the square region diagram;
the second deployment module is used for deploying the spectrum image collector along the 0 degree direction, the 45 degree direction, the 90 degree direction, the 135 degree direction, the 180 degree direction, the 225 degree direction, the 270 degree direction and the 315 degree direction of the coordinate system respectively; the method specifically comprises the following steps:
respectively disposing n spectrum image collectors along the 0 degree direction, 45 degree direction, 90 degree direction, 135 degree direction, 180 degree direction, 225 degree direction, 270 degree direction and 315 degree direction of the coordinate system in the range of the square region graph according to the principle of fixed distance average distribution;
the first feature extraction module is used for carrying out convolution operation feature extraction on the multispectral image data acquired by the first spectrum image acquisition device to obtain a feature map of the multispectral image, wherein the feature map comprises first features;
the second feature extraction module is used for respectively extracting features of n spectrum image collectors respectively deployed in 8 directions, and specifically comprises the following steps: respectively carrying out convolution operation feature extraction on the multispectral image data acquired by the n multispectral image collectors to obtain feature graphs of n multispectral images corresponding to the direction, wherein the feature graphs comprise n corresponding second features;
a first feature vector calculation module for performing difference processing operation on the n corresponding second features and the first features to obtain feature vectors p= (x) corresponding to the direction 1 ,x 2 ,...,x n )；
The second feature vector calculation module is used for repeating the steps of the second feature extraction module and the first feature vector calculation module to respectively obtain feature vectors corresponding to 8 directions;
the fusion module is used for merging and fusing the obtained 8 feature vectors corresponding to the directions to obtain a fused 8 multiplied by n feature matrix Q, whereinWherein->Representing a first eigenvector calculated along the 0 deg. direction of said coordinate system,/i>Representing an nth feature vector calculated along a 0 ° direction of the coordinate system; />Representing a first eigenvector calculated along the 315 ° direction of said coordinate system,/->Representing an nth feature vector calculated along a 315 ° direction of the coordinate system;
and the recognition module is used for inputting the fused characteristic matrix Q into a pre-trained AlexNet recognition model to obtain a gas hazard source recognition result in the target monitoring area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310466239.7A CN116189100B (en) | 2023-04-27 | 2023-04-27 | Gas hazard source detection and identification method and system based on spectral image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310466239.7A CN116189100B (en) | 2023-04-27 | 2023-04-27 | Gas hazard source detection and identification method and system based on spectral image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116189100A CN116189100A (en) | 2023-05-30 |
| CN116189100B true CN116189100B (en) | 2023-07-18 |
Family
ID=86434870
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310466239.7A Active CN116189100B (en) | 2023-04-27 | 2023-04-27 | Gas hazard source detection and identification method and system based on spectral image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116189100B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110060299A (en) * | 2019-04-18 | 2019-07-26 | 中国测绘科学研究院 | Danger source identifies and positions method in passway for transmitting electricity based on binocular vision technology |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108957573B (en) * | 2017-05-19 | 2024-04-02 | 北京英特威视科技有限公司 | Dangerous source detection device and method |
| KR101922727B1 (en) * | 2018-06-11 | 2018-11-27 | 주식회사 인더스비젼 | Gas leak detection and alarm system |
-
2023
- 2023-04-27 CN CN202310466239.7A patent/CN116189100B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110060299A (en) * | 2019-04-18 | 2019-07-26 | 中国测绘科学研究院 | Danger source identifies and positions method in passway for transmitting electricity based on binocular vision technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116189100A (en) | 2023-05-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112781791B (en) | VOCs gas leakage detection method and system based on optical gas imaging | |
| CN112529012B (en) | Oil and gas pipeline inspection method, device, equipment and medium based on target matching | |
| CN108269322B (en) | Inspection method | |
| CN110827435A (en) | External damage monitoring method and system based on intelligent warning post and readable storage medium | |
| CN110285801B (en) | Positioning method and device for intelligent safety helmet | |
| CN110766685A (en) | Power transmission line forest fire monitoring method and system based on remote sensing data cloud detection | |
| CN112185131A (en) | Vehicle driving state judgment method and device, computer equipment and storage medium | |
| US11881021B2 (en) | Method of providing carbon emission management information, and server and system using the same | |
| CN108260076B (en) | Method, platform and system for monitoring unmanned aerial vehicle running track | |
| US20210064848A1 (en) | Method and System for Prediction and Mitigation of Spontaneous Combustion in Coal Stock Piles | |
| CN108320347A (en) | A kind of robot method for inspecting | |
| CN104732103A (en) | Evaluation method for electric transmission line orientated forest fire approaching risk | |
| CN116189100B (en) | Gas hazard source detection and identification method and system based on spectral image | |
| Lei et al. | Discovering maritime traffic route from AIS network | |
| US20220083774A1 (en) | Method and system for asset inspection using unmanned aerial vehicles | |
| CN114004566A (en) | Danger warning method, device and storage medium | |
| CN107633637A (en) | A kind of power network mountain fire satellite monitoring alarm localization method based on bivariate table interpolation | |
| CN109360373B (en) | Reservoir water discharge personnel early warning method based on WIFI positioning | |
| CN204855838U (en) | Thunder and lightning positioner based on optical observation | |
| CN114299085B (en) | Dangerous source identification method based on unmanned aerial vehicle aerial photography | |
| CN107765211B (en) | System for detecting intrusion signal of unknown broadcast television | |
| CN105092783A (en) | Atmospheric pollution monitoring device, method and system, and terminal device | |
| CN114820788A (en) | Multi-parameter lightning location, system, computer device and medium | |
| CN114025375A (en) | 5G tracking area abnormity detection method and system based on honeycomb distribution identification technology | |
| CN116486307B (en) | Chemical industry park safety early warning method and system based on video analysis technology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |