CN-122023823-A - Gas emission prediction method, device, equipment, medium and product

CN122023823ACN 122023823 ACN122023823 ACN 122023823ACN-122023823-A

Abstract

The disclosure relates to the technical field of gas detection, and in particular provides a gas emission prediction method, a device, equipment, a medium and a product, which comprise the steps of acquiring a flame image of target flame; the method comprises the steps of determining a plurality of feature extraction dimensions of a flame image according to flame combustion attributes and flame combustion states of target flames, carrying out feature extraction on the flame image based on the feature extraction dimensions to obtain multi-dimensional flame feature parameters, wherein the multi-dimensional flame feature parameters are used for describing dynamic changes of the flame combustion attributes and the flame combustion states at different time points, and predicting gas emission concentration of the target flames in a combustion process according to the multi-dimensional flame feature parameters. By the processing mode, the dynamic change of flame combustion attribute and flame combustion state can be more comprehensively analyzed according to the multiple feature extraction dimensions of the target flame, and the gas emission concentration in the combustion process can be accurately predicted by utilizing the multi-dimensional flame feature parameters.

Inventors

ZHANG CHAO
ZHONG NING
KANG JUN

Assignees

国家能源集团国源电力有限公司
陕西德源府谷能源有限公司

Dates

Publication Date: 20260512
Application Date: 20260108

Claims (10)

1. A gas emission prediction method, comprising: Acquiring a flame image of a target flame; determining a plurality of feature extraction dimensions of the flame image according to the flame combustion attribute and the flame combustion state of the target flame; Performing feature extraction on the flame image based on the feature extraction dimensions to obtain multi-dimensional flame feature parameters, wherein the multi-dimensional flame feature parameters are used for describing the flame combustion attribute and the dynamic change of the flame combustion state at different time points; and predicting the gas emission concentration of the target flame in the combustion process according to the multi-dimensional flame characteristic parameters.
2. The method of claim 1, wherein the determining a plurality of feature extraction dimensions of the flame image based on flame combustion properties and flame combustion status of the target flame comprises: Determining geometric parameters and brightness parameters of the flame image according to the flame combustion attribute, wherein the geometric parameters are used for indicating the shape of the target flame, and the brightness parameters are used for indicating the combustion intensity of the target flame; Determining thermodynamic parameters and optical flow parameters of the flame image according to the flame combustion state, wherein the thermodynamic parameters are used for indicating the temperature change of the target flame, and the optical flow parameters are used for indicating the dynamic behavior change of the target flame; Determining the geometric parameter, the luminance parameter, the thermodynamic parameter, and the optical flow parameter as the plurality of feature extraction dimensions.
3. The method of claim 2, wherein said determining geometric parameters and brightness parameters of said flame image based on said flame burn attribute comprises: determining the geometric parameters according to the ignition point, root area, opening angle, flame length and flame area of the target flame; And determining the brightness parameter according to the brightness and the non-uniformity of the target flame.
4. The method of claim 2, wherein determining thermodynamic and optical flow parameters of the flame image based on the flame combustion state comprises: Determining the thermodynamic parameters according to the highest temperature, the lowest temperature, the average temperature and the flicker frequency of the target flame; And determining the optical flow parameters according to the flame optical flow speed and the direction of the target flame.
5. The method of claim 1, wherein the feature extracting the flame image based on the plurality of feature extraction dimensions results in a multi-dimensional flame feature parameter, comprising: Converting the flame image into a flame gray level image and a flame binary image; determining target images corresponding to each feature extraction dimension, wherein the target images comprise the flame images, the flame gray scale images and the flame binary images; And extracting corresponding characteristic parameters from the target image to determine the multi-dimensional flame characteristic parameters.
6. The method of claim 1, wherein predicting the gas emission concentration of the target flame during combustion based on the multi-dimensional flame signature parameter comprises: weighting the multi-dimensional flame characteristic parameters at the historical time points to obtain a weighted result; And processing the weighted result and the deviation vector through an activation function, and predicting the gas emission concentration according to the processing result.
7. A gas emission prediction apparatus, comprising: an acquisition unit for acquiring a flame image of a target flame; a dimension determining unit, configured to determine a plurality of feature extraction dimensions of the flame image according to a flame combustion attribute and a flame combustion state of the target flame; The device comprises a feature extraction unit, a multi-dimensional flame characteristic parameter, a flame combustion unit and a flame combustion unit, wherein the feature extraction unit is used for carrying out feature extraction on the flame image based on the feature extraction dimensions to obtain the multi-dimensional flame characteristic parameter, and the multi-dimensional flame characteristic parameter is used for describing the flame combustion attribute and the dynamic change of the flame combustion state at different time points; and the prediction unit is used for predicting the gas emission concentration of the target flame in the combustion process according to the multi-dimensional flame characteristic parameters.
8. An electronic device comprising a processor, a memory and a bus, the memory storing machine-readable instructions executable by the processor, the processor in communication with the memory via the bus when the electronic device is in operation, the machine-readable instructions when executed by the processor performing the steps of the gas emission prediction method of any one of claims 1 to 6.
9. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program which, when executed by a processor, performs the steps of the gas emission prediction method according to any of claims 1 to 6.
10. A computer program product, characterized in that it is stored in a storage medium, said program product being executed by at least one processor to implement the gas emission prediction method according to any one of claims 1 to 6.

Description

Gas emission prediction method, device, equipment, medium and product Technical Field The disclosure relates to the technical field of gas detection, in particular to a gas emission prediction method, a device, equipment, a medium and a product. Background At present, pollutant emission in the industrial production process has become a focus of attention, and nitrogen oxides are common harmful gases in the industrial combustion process, and the emission of the harmful gases not only causes serious harm to the environment, but also can negatively affect the health of human bodies. In the related art, complicated mathematical models and a large amount of field data support are generally required, so that the problems of low accuracy and poor instantaneity exist in the traditional method under the condition of processing large-scale data, and the real-time prediction requirement of harmful gas emission in modern industrial production cannot be met. Disclosure of Invention The present disclosure has been made in view of the above-described problems. The present disclosure provides a gas emission prediction method, apparatus, device, medium, and product. According to one aspect of the present disclosure, there is provided a gas emission prediction method including: Acquiring a flame image of a target flame; determining a plurality of feature extraction dimensions of the flame image according to the flame combustion attribute and the flame combustion state of the target flame; Performing feature extraction on the flame image based on the feature extraction dimensions to obtain multi-dimensional flame feature parameters, wherein the multi-dimensional flame feature parameters are used for describing the flame combustion attribute and the dynamic change of the flame combustion state at different time points; and predicting the gas emission concentration of the target flame in the combustion process according to the multi-dimensional flame characteristic parameters. Further, the determining a plurality of feature extraction dimensions of the flame image according to the flame combustion attribute and flame combustion state of the target flame according to an aspect of the present disclosure further includes: Determining geometric parameters and brightness parameters of the flame image according to the flame combustion attribute, wherein the geometric parameters are used for indicating the shape of the target flame, and the brightness parameters are used for indicating the combustion intensity of the target flame; Determining thermodynamic parameters and optical flow parameters of the flame image according to the flame combustion state, wherein the thermodynamic parameters are used for indicating the temperature change of the target flame, and the optical flow parameters are used for indicating the dynamic behavior change of the target flame; Determining the geometric parameter, the luminance parameter, the thermodynamic parameter, and the optical flow parameter as the plurality of feature extraction dimensions. Further, the determining geometrical parameters and brightness parameters of the flame image according to the flame combustion attribute according to an aspect of the present disclosure further includes: determining the geometric parameters according to the ignition point, root area, opening angle, flame length and flame area of the target flame; And determining the brightness parameter according to the brightness and the non-uniformity of the target flame. Further, the determining thermodynamic and optical flow parameters of the flame image according to the flame combustion state according to an aspect of the present disclosure further includes: Determining the thermodynamic parameters according to the highest temperature, the lowest temperature, the average temperature and the flicker frequency of the target flame; And determining the optical flow parameters according to the flame optical flow speed and the direction of the target flame. In addition, the feature extraction of the flame image based on the plurality of feature extraction dimensions according to an aspect of the disclosure obtains a multi-dimensional flame feature parameter, and the method further includes: Converting the flame image into a flame gray level image and a flame binary image; determining target images corresponding to each feature extraction dimension, wherein the target images comprise the flame images, the flame gray scale images and the flame binary images; And extracting corresponding characteristic parameters from the target image to determine the multi-dimensional flame characteristic parameters. Further, the predicting a gas emission concentration of the target flame during combustion according to the multi-dimensional flame characteristic parameter according to an aspect of the present disclosure further comprises: weighting the multi-dimensional flame characteristic parameters at the historical time points to obtain a weig