CN-121577539-B - Optical power self-adaptive adjusting method and system for laser methane detector

CN121577539BCN 121577539 BCN121577539 BCN 121577539BCN-121577539-B

Abstract

The application provides a self-adaptive adjustment method and a self-adaptive adjustment system for optical power of a laser methane detector, which relate to the technical field of gas detection, wherein the method comprises the steps of collecting ambient light data, target distance information and reflected laser signals and calculating signal-to-noise ratio data; the method comprises the steps of inputting environment light data, target distance information and signal to noise ratio data into a convolutional neural network to perform feature extraction to obtain multi-dimensional environment features, inputting the multi-dimensional environment features into a fuzzy logic processor to perform scene recognition and output scene recognition results, inquiring a preset mapping table according to the scene recognition results to obtain a driving current target value, determining a current adjustment quantity according to the current actual value, and adjusting driving current of a transmitting module accordingly, so that the self-adaptive adjustment of the optical power of the laser methane detector is realized. The application improves the detection stability and adaptability of the laser methane detector in different complex environments.

Inventors

ZHAO WANYI
LI LIN
YIN ZHEN
WANG LIHAN

Assignees

天津市极光创新智能科技有限公司
天津市极光光电科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260121

Claims (7)

1. A method for adaptively adjusting optical power of a laser methane detector, comprising: Collecting ambient light data, target distance information and reflected laser signals; Calculating signal-to-noise ratio data based on the ambient light data and the reflected laser signal; Inputting the ambient light data, the target distance information and the signal to noise ratio data into a convolutional neural network for feature extraction to obtain multi-dimensional environmental features; Inputting the multi-dimensional environmental characteristics to a fuzzy logic processor for scene recognition, and outputting a scene recognition result; inquiring a preset mapping table according to the scene recognition result to obtain a target value of the driving current of a transmitting module in the laser methane detector, and determining a current adjustment quantity according to the target value and the actual value of the driving current; Adjusting the driving current of the emission module based on the current adjustment amount to realize the self-adaptive adjustment of the optical power of the laser methane detector; inputting the multi-dimensional environmental features to a fuzzy logic processor for scene recognition, and outputting scene recognition results, wherein the method comprises the following steps: Inputting the multi-dimensional environmental characteristics to a fuzzification interface of a fuzzy logic processor, and respectively converting the light intensity distribution characteristics, the distance change characteristics and the signal quality characteristics in the multi-dimensional environmental characteristics into corresponding fuzzy sets through a plurality of membership functions preset in the fuzzification interface; inputting all fuzzy sets into a rule base of the fuzzy logic processor, wherein the rule base comprises a plurality of preset reasoning rules, carrying out reasoning operation on all fuzzy sets according to the reasoning rules to obtain fuzzy output results, and outputting scene recognition results based on the fuzzy output results; and carrying out reasoning operation on all fuzzy sets according to the reasoning rule to obtain a fuzzy output result, and outputting a scene recognition result based on the fuzzy output result, wherein the method comprises the following steps of: Calculating the confidence coefficient of the corresponding conclusion part according to the satisfaction degree of the fuzzy set on the corresponding condition part in the reasoning rule; combining the confidence degrees of the conclusion parts corresponding to all the fuzzy sets to generate a fuzzy output result; based on the fuzzy output result, adopting an improved gravity center method to respectively calculate a plurality of determined values corresponding to a plurality of scene categories to form a determined value sequence; Inputting the determined value sequence into a gating circulation unit network for time sequence characteristic enhancement, and generating an enhanced determined value sequence; Based on the enhanced determined value sequence, classifying the scene by using a decision tree model to generate a scene recognition result; the step of generating a scene recognition result by classifying the enhanced determined value sequence through a decision tree model comprises the following steps: inputting the enhanced determined value sequence into a decision tree model, wherein the decision tree model comprises a root node, a plurality of branch nodes and a plurality of leaf nodes, each branch node corresponds to an attribute judgment condition, and each leaf node corresponds to a scene category; The decision tree model sequentially judges whether the attribute judging condition of the branch node is met according to the enhanced determined value sequence from the root node, and traverses to a leaf node along a path meeting the condition; and outputting the scene category corresponding to the traversed leaf node as a scene recognition result.
2. The method of claim 1, wherein the calculating signal-to-noise ratio data based on the ambient light data and the reflected laser signal comprises: Analyzing the reflected laser signal to obtain a first component and a second component; Extracting background light data at corresponding moments from the ambient light data; Inputting the first component and the background light data into an adaptive filtering model, calculating effective signal data through the adaptive filtering model, and taking the second component as noise data; And inputting the effective signal data and the noise data into an evaluation network for calculation to obtain signal-to-noise ratio data.
3. The method of claim 1, wherein inputting the ambient light data, the target distance information, and the signal-to-noise ratio data into a convolutional neural network for feature extraction to obtain a multi-dimensional ambient feature comprises: inputting the ambient light data, the target distance information and the signal to noise ratio data into a first branch, a second branch and a third branch of the convolutional neural network respectively for processing; In the first branch, performing a first convolution operation on the ambient light data to extract light intensity distribution characteristics, in the second branch, performing a second convolution operation on the target distance information to extract distance change characteristics, and in the third branch, performing a third convolution operation on the signal-to-noise ratio data to extract signal quality characteristics; and combining the light intensity distribution characteristic, the distance change characteristic and the signal quality characteristic to generate a multi-dimensional environment characteristic.
4. The method according to claim 1, wherein the step of inquiring a preset mapping table according to the scene recognition result to obtain a target value of the driving current of the emission module in the laser methane detector, and determining the current adjustment amount according to the target value and the actual value of the driving current comprises: Searching a corresponding driving current set value in the mapping table according to the scene recognition result, and taking the driving current set value as a target value; Collecting the current value of the driving current of the transmitting module as an actual value; and calculating a difference value between the target value and the actual value, and taking the difference value as a current adjustment quantity.
5. An optical power adaptive adjustment system for a laser methane detector, comprising: the acquisition module is used for acquiring the ambient light data, the target distance information and the reflected laser signals; The calculating module is used for calculating signal-to-noise ratio data based on the ambient light data and the reflected laser signals; the extraction module is used for inputting the ambient light data, the target distance information and the signal-to-noise ratio data into a convolutional neural network to perform feature extraction so as to obtain multi-dimensional ambient features; The recognition module is used for inputting the multi-dimensional environmental characteristics to the fuzzy logic processor for scene recognition and outputting scene recognition results; the determining module is used for inquiring a preset mapping table according to the scene identification result, obtaining a target value of the driving current of the transmitting module in the laser methane detector, and determining a current adjustment quantity according to the target value and the actual value of the driving current; the adjusting module is used for adjusting the driving current of the transmitting module based on the current adjusting quantity so as to realize the self-adaptive adjustment of the optical power of the laser methane detector; inputting the multi-dimensional environmental features to a fuzzy logic processor for scene recognition, and outputting scene recognition results, wherein the method comprises the following steps: Inputting the multi-dimensional environmental characteristics to a fuzzification interface of a fuzzy logic processor, and respectively converting the light intensity distribution characteristics, the distance change characteristics and the signal quality characteristics in the multi-dimensional environmental characteristics into corresponding fuzzy sets through a plurality of membership functions preset in the fuzzification interface; inputting all fuzzy sets into a rule base of the fuzzy logic processor, wherein the rule base comprises a plurality of preset reasoning rules, carrying out reasoning operation on all fuzzy sets according to the reasoning rules to obtain fuzzy output results, and outputting scene recognition results based on the fuzzy output results; and carrying out reasoning operation on all fuzzy sets according to the reasoning rule to obtain a fuzzy output result, and outputting a scene recognition result based on the fuzzy output result, wherein the method comprises the following steps of: Calculating the confidence coefficient of the corresponding conclusion part according to the satisfaction degree of the fuzzy set on the corresponding condition part in the reasoning rule; combining the confidence degrees of the conclusion parts corresponding to all the fuzzy sets to generate a fuzzy output result; based on the fuzzy output result, adopting an improved gravity center method to respectively calculate a plurality of determined values corresponding to a plurality of scene categories to form a determined value sequence; Inputting the determined value sequence into a gating circulation unit network for time sequence characteristic enhancement, and generating an enhanced determined value sequence; Based on the enhanced determined value sequence, classifying the scene by using a decision tree model to generate a scene recognition result; the step of generating a scene recognition result by classifying the enhanced determined value sequence through a decision tree model comprises the following steps: inputting the enhanced determined value sequence into a decision tree model, wherein the decision tree model comprises a root node, a plurality of branch nodes and a plurality of leaf nodes, each branch node corresponds to an attribute judgment condition, and each leaf node corresponds to a scene category; The decision tree model sequentially judges whether the attribute judging condition of the branch node is met according to the enhanced determined value sequence from the root node, and traverses to a leaf node along a path meeting the condition; and outputting the scene category corresponding to the traversed leaf node as a scene recognition result.
6. An electronic device, comprising: A memory for storing a computer program; A processor for implementing the steps of the method for adaptively adjusting the optical power of a laser methane detector according to any one of claims 1 to 4 when said computer program is executed.
7. A computer-readable storage medium, in which a computer program is stored, which computer program, when being executed by a processor, is capable of implementing a method for adaptively adjusting the optical power of a laser methane detector according to any one of claims 1 to 4.

Description

Optical power self-adaptive adjusting method and system for laser methane detector Technical Field The application relates to the technical field of gas detection, in particular to a self-adaptive adjustment method and system for optical power of a laser methane detector. Background The laser methane detector plays an important role in safety monitoring in the industrial fields of chemical industry, gas industry and the like, recognizes leakage by detecting the absorption of methane gas to laser with specific wavelength, has the advantages of high sensitivity and quick response, and becomes key how to maintain the stability of detection signals along with the expansion of application scenes from open areas to complex environments such as tunnels, pipelines, indoor corridor and the like. In the prior art, there are methods for controlling the optical power by collecting single ambient light intensity information and adjusting the output power of the laser based on a preset threshold value, and combining the ambient light sensor with the distance measuring unit through a fixed mapping relation. These methods can meet the basic requirements when the environmental conditions change smoothly. However, when the above-mentioned existing method faces complex and changeable on-site working conditions, for example, when the reflected light of the environment is strong, the distance between the detection targets dynamically changes, and the background noise is complex, the response of the power adjustment is not accurate enough and timely, which results in fluctuation of the detection signal-to-noise ratio, and affects the reliability and stability of the final detection result. Therefore, the prior art has the technical problem of insufficient adaptability to complex working conditions. Disclosure of Invention The application provides a self-adaptive adjustment method and a self-adaptive adjustment system for optical power of a laser methane detector, which are used for solving the problems of poor detection stability and poor adaptability of the laser methane detector in different complex environments in the prior art. In order to solve the above technical problems, in a first aspect, the present application provides a method for adaptively adjusting optical power of a laser methane detector, including: Collecting ambient light data, target distance information and reflected laser signals; Calculating signal-to-noise ratio data based on the ambient light data and the reflected laser signal; Inputting the ambient light data, the target distance information and the signal to noise ratio data into a convolutional neural network for feature extraction to obtain multi-dimensional environmental features; Inputting the multi-dimensional environmental characteristics to a fuzzy logic processor for scene recognition, and outputting a scene recognition result; inquiring a preset mapping table according to the scene recognition result to obtain a target value of the driving current of a transmitting module in the laser methane detector, and determining a current adjustment quantity according to the target value and the actual value of the driving current; and adjusting the driving current of the emission module based on the current adjustment quantity so as to realize the self-adaptive adjustment of the optical power of the laser methane detector. Optionally, the inputting the multi-dimensional environmental feature to a fuzzy logic processor for scene recognition, outputting a scene recognition result, includes: Inputting the multi-dimensional environmental characteristics to a fuzzification interface of a fuzzy logic processor, and respectively converting the light intensity distribution characteristics, the distance change characteristics and the signal quality characteristics in the multi-dimensional environmental characteristics into corresponding fuzzy sets through a plurality of membership functions preset in the fuzzification interface; inputting all fuzzy sets into a rule base of the fuzzy logic processor, wherein the rule base comprises a plurality of preset reasoning rules, carrying out reasoning operation on all fuzzy sets according to the reasoning rules to obtain fuzzy output results, and outputting scene recognition results based on the fuzzy output results. Optionally, performing an inference operation on all fuzzy sets according to the inference rule to obtain a fuzzy output result, and outputting a scene recognition result based on the fuzzy output result, including: Calculating the confidence coefficient of the corresponding conclusion part according to the satisfaction degree of the fuzzy set on the corresponding condition part in the reasoning rule; combining the confidence degrees of the conclusion parts corresponding to all the fuzzy sets to generate a fuzzy output result; based on the fuzzy output result, adopting an improved gravity center method to respectively calculate a plurality of determined valu