CN117147532B - Micro-chemiluminescence method nitrogen oxide gas sensor based on Internet of things - Google Patents
Micro-chemiluminescence method nitrogen oxide gas sensor based on Internet of things Download PDFInfo
- Publication number
- CN117147532B CN117147532B CN202311144880.5A CN202311144880A CN117147532B CN 117147532 B CN117147532 B CN 117147532B CN 202311144880 A CN202311144880 A CN 202311144880A CN 117147532 B CN117147532 B CN 117147532B
- Authority
- CN
- China
- Prior art keywords
- electric signal
- waveform
- training
- feature
- signal waveform
- 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
- MWUXSHHQAYIFBG-UHFFFAOYSA-N Nitric oxide Chemical compound O=[N] MWUXSHHQAYIFBG-UHFFFAOYSA-N 0.000 title claims abstract description 187
- 239000007789 gas Substances 0.000 title claims abstract description 70
- 238000000034 method Methods 0.000 title claims abstract description 37
- 230000003287 optical effect Effects 0.000 claims abstract description 21
- 230000003321 amplification Effects 0.000 claims abstract description 18
- 238000003199 nucleic acid amplification method Methods 0.000 claims abstract description 18
- 238000012549 training Methods 0.000 claims description 73
- 238000013527 convolutional neural network Methods 0.000 claims description 34
- 238000000605 extraction Methods 0.000 claims description 32
- 238000003062 neural network model Methods 0.000 claims description 24
- 238000001514 detection method Methods 0.000 claims description 21
- 230000008569 process Effects 0.000 claims description 19
- 238000010586 diagram Methods 0.000 claims description 15
- 230000014509 gene expression Effects 0.000 claims description 14
- 230000004913 activation Effects 0.000 claims description 13
- 230000004927 fusion Effects 0.000 claims description 12
- 238000000265 homogenisation Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims 2
- 238000012545 processing Methods 0.000 abstract description 15
- 238000005259 measurement Methods 0.000 abstract description 13
- 238000004458 analytical method Methods 0.000 abstract description 9
- 238000004422 calculation algorithm Methods 0.000 abstract description 9
- 230000035945 sensitivity Effects 0.000 abstract description 2
- 230000006855 networking Effects 0.000 abstract 1
- 230000003014 reinforcing effect Effects 0.000 abstract 1
- 238000011176 pooling Methods 0.000 description 13
- 239000000306 component Substances 0.000 description 8
- 239000011159 matrix material Substances 0.000 description 8
- 238000013528 artificial neural network Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 5
- 238000013507 mapping Methods 0.000 description 5
- RAHZWNYVWXNFOC-UHFFFAOYSA-N Sulphur dioxide Chemical compound O=S=O RAHZWNYVWXNFOC-UHFFFAOYSA-N 0.000 description 4
- 239000003153 chemical reaction reagent Substances 0.000 description 4
- 238000013136 deep learning model Methods 0.000 description 4
- 230000007613 environmental effect Effects 0.000 description 4
- 238000005070 sampling Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- CBENFWSGALASAD-UHFFFAOYSA-N Ozone Chemical compound [O-][O+]=O CBENFWSGALASAD-UHFFFAOYSA-N 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000003066 decision tree Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000004377 microelectronic Methods 0.000 description 2
- 238000003909 pattern recognition Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000003213 activating effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000003750 conditioning effect Effects 0.000 description 1
- 239000008358 core component Substances 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000003344 environmental pollutant Substances 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 231100000719 pollutant Toxicity 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 230000008054 signal transmission Effects 0.000 description 1
- 238000000547 structure data Methods 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 239000011800 void material Substances 0.000 description 1
Classifications
-
- 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
- G01N21/75—Systems in which material is subjected to a chemical reaction, the progress or the result of the reaction being investigated
- G01N21/76—Chemiluminescence; Bioluminescence
-
- 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
- G01N21/01—Arrangements or apparatus for facilitating the optical investigation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A50/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE in human health protection, e.g. against extreme weather
- Y02A50/20—Air quality improvement or preservation, e.g. vehicle emission control or emission reduction by using catalytic converters
Landscapes
- Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Plasma & Fusion (AREA)
- Engineering & Computer Science (AREA)
- Investigating Or Analysing Materials By The Use Of Chemical Reactions (AREA)
Abstract
The application discloses miniature chemiluminescence method nitrogen oxide gas sensor based on thing networking, it is after the optical signal that the sensor measured through photoelectric detector converts into the electrical signal, amplify the processing to the electrical signal, and introduce signal processing and analysis algorithm at the rear end and realize the waveform analysis to the electrical signal after amplifying, so the waveform characteristic information that has the nitrogen oxide gas in the electrical signal after catching the amplification, thereby reduce the influence of noise to weak electrical signal, the interference killing feature of reinforcing sensor realizes the high accuracy, the high sensitivity measurement to nitrogen oxide gas concentration.
Description
Technical Field
The application relates to the field of intelligent detection, and more particularly, to a micro-chemiluminescence method nitrogen oxide gas sensor based on the Internet of things.
Background
Nitrogen oxides (NOx) are an important class of atmospheric pollutants that have serious environmental and human health implications. Therefore, developing a high-precision and high-sensitivity nitrogen oxide gas sensor has important significance for monitoring and controlling the atmospheric pollution.
Conventional nitrogen oxide gas sensors are typically large and complex, requiring the use of a large number of optical devices, reagents and electronic components, which makes the sensor bulky and inconvenient to integrate and apply to miniaturized systems. The micro-chemiluminescence nitrogen oxide gas sensor adopts a micro-design, combines a micro-fluidic technology and a micro-electronic processing technology, and has the characteristics of small size, low power consumption and quick response.
In the micro-chemiluminescence method nitrogen oxide gas sensor, nitrogen oxide gas and specific reagents (such as ozone, sulfur dioxide and the like) are subjected to chemical reaction in the sensor, and the generated excited molecules emit visible light chemiluminescence signals after a de-excitation process. The optical signal received by the sensor is then converted into a weak electrical signal by the photodetector. Due to weak signals, they are susceptible to environmental interference and electronic noise, which can lead to reduced signal quality and reduced measurement accuracy. In order to enhance the signal strength, it is generally necessary to amplify the received electrical signal. However, in the amplification process, not only useful signals but also noise signals are amplified, so that noise becomes a main limiting factor in the whole system, the accuracy of measurement is further affected, and accurate measurement results are difficult to obtain by the nitrogen oxide gas sensor.
Therefore, an optimized micro-chemiluminescent nitrogen oxide gas sensor based on the Internet of things is desired.
Disclosure of Invention
The present application has been made in order to solve the above technical problems. The embodiment of the application provides a miniature chemiluminescence method nitrogen oxide gas sensor based on the Internet of things, which is characterized in that after an optical signal measured by the sensor is converted into an electric signal through a photoelectric detector, the electric signal is amplified, and a signal processing and analyzing algorithm is introduced into the rear end to realize waveform analysis of the amplified electric signal, so that waveform characteristic information about nitrogen oxide gas in the amplified electric signal is captured, the influence of noise on the weak electric signal is reduced, the anti-interference capability of the sensor is enhanced, and the high-precision and high-sensitivity measurement of the concentration of the nitrogen oxide gas is realized.
According to one aspect of the present application, there is provided a micro chemiluminescent nitrogen oxide gas sensor based on the internet of things, comprising:
the data signal acquisition module is used for acquiring the electric signals acquired by the optical detection system;
the signal amplifying module is used for amplifying the electric signal to obtain an amplified electric signal;
The electric signal waveform characteristic extraction module is used for extracting waveform characteristics of the amplified electric signal to obtain electric signal waveform characteristics;
and the nitrogen oxide gas concentration detection module is used for determining the concentration value of the nitrogen oxide gas based on the waveform characteristics of the electric signals.
Compared with the prior art, the micro-chemiluminescence method nitrogen oxide gas sensor based on the Internet of things has the advantages that after the photoelectric detector converts the optical signal measured by the sensor into the electric signal, the electric signal is amplified, and the waveform analysis of the amplified electric signal is realized by introducing the signal processing and analysis algorithm at the rear end, so that waveform characteristic information about nitrogen oxide gas in the amplified electric signal is captured, the influence of noise on the weak electric signal is reduced, the anti-interference capability of the sensor is enhanced, and the high-precision and high-sensitivity measurement of the concentration of the nitrogen oxide gas is realized.
Drawings
The foregoing and other objects, features and advantages of the present application will become more apparent from the following more particular description of embodiments of the present application, as illustrated in the accompanying drawings. The accompanying drawings are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate the application and not constitute a limitation to the application. In the drawings, like reference numerals generally refer to like parts or steps.
FIG. 1 is a block diagram of a micro-chemiluminescent nitrogen oxide gas sensor based on the Internet of things in accordance with an embodiment of the present application;
FIG. 2 is a system architecture diagram of a micro-chemiluminescent process NOx gas sensor based on the Internet of things in accordance with an embodiment of the present application;
FIG. 3 is a flow chart of a training phase of a micro-chemiluminescent nitrogen oxide gas sensor based on the Internet of things in accordance with an embodiment of the present application;
fig. 4 is a block diagram of an electrical signal waveform feature extraction module in a micro-chemiluminescence method nitrogen oxide gas sensor based on the internet of things according to an embodiment of the application.
Detailed Description
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application and not all of the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
As used in this application and in the claims, the terms "a," "an," "the," and/or "the" are not specific to the singular, but may include the plural, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.
Although the present application makes various references to certain modules in a system according to embodiments of the present application, any number of different modules may be used and run on a user terminal and/or server. The modules are merely illustrative, and different aspects of the systems and methods may use different modules.
Flowcharts are used in this application to describe the operations performed by systems according to embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in order precisely. Rather, the various steps may be processed in reverse order or simultaneously, as desired. Also, other operations may be added to or removed from these processes.
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application and not all of the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
Conventional nitrogen oxide gas sensors are typically large and complex, requiring the use of a large number of optical devices, reagents and electronic components, which makes the sensor bulky and inconvenient to integrate and apply to miniaturized systems. The micro-chemiluminescence nitrogen oxide gas sensor adopts a micro-design, combines a micro-fluidic technology and a micro-electronic processing technology, and has the characteristics of small size, low power consumption and quick response. In the micro-chemiluminescence method nitrogen oxide gas sensor, nitrogen oxide gas and specific reagents (such as ozone, sulfur dioxide and the like) are subjected to chemical reaction in the sensor, and the generated excited molecules emit visible light chemiluminescence signals after a de-excitation process. The optical signal received by the sensor is then converted into a weak electrical signal by the photodetector. Due to weak signals, they are susceptible to environmental interference and electronic noise, which can lead to reduced signal quality and reduced measurement accuracy. In order to enhance the signal strength, it is generally necessary to amplify the received electrical signal. However, in the amplification process, not only useful signals but also noise signals are amplified, so that noise becomes a main limiting factor in the whole system, the accuracy of measurement is further affected, and accurate measurement results are difficult to obtain by the nitrogen oxide gas sensor. Therefore, an optimized micro-chemiluminescent nitrogen oxide gas sensor based on the Internet of things is desired.
In the technical scheme of the application, a miniature chemiluminescent method nitrogen oxide gas sensor based on the Internet of things is provided. Fig. 1 is a block diagram of a micro-chemiluminescent nitrogen oxide gas sensor based on the internet of things according to an embodiment of the application. Fig. 2 is a system architecture diagram of a micro-chemiluminescent oxynitride gas sensor based on the internet of things according to an embodiment of the present application. As shown in fig. 1 and 2, a micro chemiluminescent nitrogen oxide gas sensor 300 based on the internet of things according to an embodiment of the present application comprises: a data signal acquisition module 310 for acquiring the electrical signal acquired by the optical detection system; a signal amplifying module 320, configured to amplify the electrical signal to obtain an amplified electrical signal; an electrical signal waveform feature extraction module 330, configured to perform waveform feature extraction on the amplified electrical signal to obtain an electrical signal waveform feature; the nox gas concentration detection module 340 is configured to determine a concentration value of the nox gas based on the waveform characteristic of the electrical signal.
In particular, the data signal acquisition module 310 is configured to acquire the electrical signal acquired by the optical detection system. It should be noted that an optical detection system is a system that performs detection and measurement using optical principles and techniques. The device uses optical sensor, light source, optical element and image process to analyze and process the light character, to realize the measurement and judge of the character, shape, position, size and color parameters of the detected object.
Accordingly, in one possible implementation, the electrical signal acquired by the optical detection system may be acquired by, for example: the optical sensor is typically constituted by a photosensitive element, such as a Photodiode (photo diode) or a photo converter (photo sensor). When light impinges on the photosensitive element, it produces a current or voltage signal; because the electrical signal generated by the photosensitive element is weak, the signal is usually amplified by an amplifier to increase the amplitude and reliability of the signal; in some cases, the photosensitive element may be affected by noise or interference, and thus a filtering process of the signal is required. The filter can remove unnecessary frequency components so as to improve the quality and stability of signals; depending on the particular application requirements, further conditioning of the signal may be required, such as gain adjustment, bias calibration, linearization, etc.; the analog signal is converted to a digital signal, which is typically sampled and quantized using an analog-to-digital converter (ADC). This converts the continuous signal acquired by the optical detection system into a discrete digital signal for subsequent digital signal processing and analysis; the digital signal is processed and analyzed by an image processing algorithm, a signal processing algorithm, or other specific analysis algorithm. This may include extracting features, measuring parameters, performing image reconstruction, performing pattern recognition, etc.
In particular, the signal amplifying module 320 is configured to amplify the electrical signal to obtain an amplified electrical signal. It will be appreciated that the electrical signal produced by the optical detection system is typically weaker and the amplification process may increase the amplitude of the signal so that it is easier to detect and analyze. The amplified signal can improve the sensitivity and reliability of the system. In addition, various noise or interference may be introduced during signal transmission and processing. The amplification process may amplify the amplitude of the signal above a noise level, thereby improving the signal-to-noise ratio. This helps to reduce the effect of noise on the signal and improves the measurement accuracy and precision of the system.
Accordingly, in one possible implementation, the electrical signal may be amplified to obtain an amplified electrical signal by, for example: depending on the characteristics and requirements of the signal, an appropriate amplifier is selected. The amplifier may be an analog amplifier or a digital amplifier, the specific choice depending on the type of signal and processing requirements; connecting the collected electric signals to the input end of the amplifier through a cable or a connector; the gain of the amplifier is set according to the amplitude range and the need of the signal. The gain determines the multiple by which the signal is amplified in the amplifier; the amplifier is powered on and is enabled to start working. The amplifier amplifies the collected electric signals to increase the amplitude of the signals; the amplified signal is connected to a subsequent processing device or system by a cable or connector.
In particular, the electrical signal waveform feature extraction module 330 is configured to perform waveform feature extraction on the amplified electrical signal to obtain an electrical signal waveform feature. In particular, in one specific example of the present application, as shown in fig. 4, the electrical signal waveform feature extraction module 330 includes: the waveform shallow feature extraction unit 331 is configured to perform shallow feature extraction on the amplified electrical signal by using a waveform shallow feature extractor based on the first deep neural network model to obtain an electrical signal waveform shallow feature map; a waveform deep feature extraction unit 332, configured to perform deep feature extraction on the electrical signal waveform shallow feature map by using a waveform deep feature extractor based on a second deep neural network model, so as to obtain an electrical signal waveform deep feature map; a joint semantic propagation unit 333, configured to use a joint semantic propagation module to fuse the electrical signal waveform deep feature map and the electrical signal waveform shallow feature map to obtain a fused semantic electrical signal waveform shallow feature map; and the receptive field amplifying unit 334 is configured to perform receptive field amplification on the fused semantic electric signal waveform shallow feature map to obtain a global semantic electric signal waveform feature map as the electric signal waveform feature.
Specifically, the waveform shallow feature extraction unit 331 is configured to perform shallow feature extraction on the amplified electrical signal by using a waveform shallow feature extractor based on a first deep neural network model to obtain a waveform shallow feature map of the electrical signal. In consideration of the fact that shallow semantic feature information of the electric signal is required to be focused more in the process of carrying out feature analysis on the amplified electric signal to inhibit noise and expression of irrelevant information, useful components of the electric signal are enhanced, and signal-to-noise ratio of the signal and noise is improved. Therefore, in the technical scheme of the application, the amplified electric signal is further subjected to feature mining in a waveform shallow feature extractor based on the first convolutional neural network model so as to extract shallow semantic feature information of the amplified electric signal, and therefore an electric signal waveform shallow feature map is obtained. In particular, the waveform shallow feature extractor can automatically learn relevant shallow semantic features in the amplified electrical signal, which enables the system to adapt to different types of signals and changing environmental conditions, improving the adaptability and generalization capability of the sensor. Specifically, each layer of the waveform shallow layer feature extractor based on the first convolutional neural network model is used for respectively carrying out input data in forward transfer of the layer: carrying out convolution processing on input data to obtain a convolution characteristic diagram; pooling the convolution feature images based on the local feature matrix to obtain pooled feature images; performing nonlinear activation on the pooled feature map to obtain an activated feature map; the output of the last layer of the waveform shallow feature extractor based on the first convolutional neural network model is the waveform shallow feature map of the electric signal, and the input of the first layer of the waveform shallow feature extractor based on the first convolutional neural network model is the amplified electric signal.
Notably, convolutional neural networks (Convolutional Neural Network, CNN) are a deep learning model that is primarily used to process tasks with grid structure data, such as images and video. The core idea of CNN is to learn and extract the features of the input data through the convolutional layer, pooling layer and fully-connected layer. The following are the main components of CNN: convolution layer: the convolutional layer is one of the core components of the CNN. It convolves the input data with a set of learnable filters (also called convolution kernels). The convolution operation performs local feature extraction on the input data through a sliding filter, thereby capturing spatial structure information of the input data. Each filter can generate a corresponding characteristic map, and a plurality of filters can extract a plurality of different characteristics; pooling layer: the pooling layer is used for reducing the space size of the feature map and reducing the number of parameters, thereby reducing the calculation amount and the memory requirement. Common pooling operations include maximum pooling and average pooling, which extract the maximum or average of local regions in the feature map as the pooling result, respectively; full tie layer: and the full connection layer flattens the feature map output by the pooling layer into a one-dimensional vector and is connected with the weights to form a full connection structure. The full connection layer is used for classifying or regressing the extracted features; activation function: between the various layers of the CNN, an activation function, such as a ReLU, is typically inserted for introducing nonlinear transformations that increase the expressive power of the network. CNNs learn progressively the abstract feature representation of the input data through multi-layer convolution, pooling, and stacking of fully connected layers. During training, the CNN automatically adjusts network parameters via a back-propagation algorithm to minimize the gap between the predicted output and the real label. This enables CNNs to achieve excellent performance in computer vision tasks such as image classification, object detection, image generation, and the like.
Specifically, the waveform deep feature extraction unit 332 is configured to perform deep feature extraction on the electrical signal waveform shallow feature map by using a waveform deep feature extractor based on a second deep neural network model to obtain an electrical signal waveform deep feature map. That is, in the technical solution of the present application, in order to further extract the higher-level characteristic information related to the amplified electrical signal, so as to describe the characteristics of the signal more comprehensively and accurately, in the technical solution of the present application, it is necessary to further pass the electrical signal waveform shallow layer feature map through a waveform deep layer feature extractor based on a second convolutional neural network model so as to obtain the electrical signal waveform deep layer feature map. It should be appreciated that the waveform deep-level feature extractor is capable of learning and extracting more abstract, higher-level features about the electrical signal, which may include timing relationships, frequency patterns, periodic variations, and the like. By capturing these advanced features, the complexity and dynamics of the signal can be better described. In addition, the waveform deep feature extractor can fuse the features of different layers, so that information of multiple layers is comprehensively considered, the expressive power and the distinguishing degree of the features are improved, and the system can better distinguish different signal modes and noise. Specifically, each layer of the waveform deep feature extractor based on the second deep neural network model is used for respectively carrying out input data in forward transfer of the layer: carrying out convolution processing on input data to obtain a convolution characteristic diagram; pooling the convolution feature images based on the local feature matrix to obtain pooled feature images; performing nonlinear activation on the pooled feature map to obtain an activated feature map; the output of the last layer of the waveform deep feature extractor based on the second deep neural network model is the electric signal waveform deep feature map, and the input of the first layer of the waveform deep feature extractor based on the second deep neural network model is the electric signal waveform shallow feature map.
Specifically, the joint semantic propagation unit 333 is configured to use a joint semantic propagation module to fuse the electrical signal waveform deep feature map and the electrical signal waveform shallow feature map to obtain a fused semantic electrical signal waveform shallow feature map. That is, a joint semantic propagation module is used to fuse the electrical signal waveform deep feature map and the electrical signal waveform shallow feature map to obtain a fused semantic electrical signal waveform shallow feature map. In particular, the joint semantic propagation module can be used for fusing semantic information in the deep feature map of the electric signal waveform into the shallow feature map of the electric signal waveform through learning and propagation, so that each position in the feature map can benefit from the semantic information of other positions, the consistency and the expression capacity of the features are improved, the fused feature map has richer semantic information, the semantic understanding and the feature expression capacity of the amplified electric signal are improved, and the accuracy of detecting the concentration of the nitrogen oxide gas is improved.
Accordingly, in one possible implementation, the electrical signal waveform deep feature map and the electrical signal waveform shallow feature map may be fused using a joint semantic propagation module to obtain a fused semantic electrical signal waveform shallow feature map, for example: first, a deep feature map and a shallow feature map are extracted from an electrical signal waveform. This may be achieved by feature extraction of the electrical signal using a deep learning model such as Convolutional Neural Network (CNN). Deep feature maps typically contain more abstract and high-level feature representations, while shallow feature maps contain lower-level and local feature representations; the joint semantic propagation module is a component for fusing deep feature maps and shallow feature maps. It may take different forms such as attention mechanisms, feature fusion operations, etc. The module aims to combine the deep feature map and the shallow feature map through effective information transmission and fusion so as to acquire richer semantic information; since the deep and shallow feature maps may differ in spatial dimensions, a feature map alignment operation is required. This can be achieved by interpolation, clipping or padding, etc. to ensure that the two feature maps have the same spatial dimensions; and inputting the aligned deep feature map and shallow feature map into a joint semantic propagation module for feature fusion. The specific fusion mode can be selected according to actual requirements, such as weighting and fusing important information of different feature maps by using an attention mechanism; and obtaining a fused semantic electric signal waveform shallow feature map after feature fusion. This fused feature map will contain rich semantic information from the deep feature map and the shallow feature map.
Specifically, the receptive field amplifying unit 334 is configured to perform receptive field amplification on the shallow feature map of the fused semantic electric signal waveform to obtain a global semantic electric signal waveform feature map as the electric signal waveform feature. Considering that convolution is a typical local operation, it can only extract local features of the waveform diagram of the electrical signal, but cannot pay attention to the global, and can affect the detection accuracy. In order to further expand the receptive field of the implicit feature expression of the electrical signal, capturing wider context information to obtain a more global and semantic feature expression, in particular, in a specific example of the application, the fused semantic electrical signal waveform shallow feature map is further passed through a feature receptive field amplifier based on a non-local neural network model to obtain a global semantic electrical signal waveform feature map. In particular, the feature receptive field amplimers are herein capable of capturing a wider range of contextual information, including global semantic relationships and long range dependencies, which helps to improve the semantic expressive power of features so that the overall structure and semantic meaning of the electrical signal waveforms can be better understood by the network. Moreover, the non-local operation can capture the association relation between remote positions, not only limited to local neighborhood, which is helpful to process long-range dependence and global consistency in signals and improve the consistency and stability of features.
Notably, the Non-local neural network (Non-Local Neural Network) is a deep learning model for computer vision tasks, aimed at capturing long-range dependencies in images or videos. In contrast to conventional Convolutional Neural Networks (CNNs), non-local neural networks introduce non-local operations that enable the network to interact and model information in a global scope. In conventional CNNs, the convolution operation is local, with each convolution kernel focusing on only a small region of the input image. This local nature is effective in many tasks, but for some tasks that require global context information, such as image segmentation, object tracking, and video analysis, local operations may not capture long-range dependencies. Non-local neural networks address this problem by introducing non-local operations. The non-local operation establishes a link between global contexts by computing similarities between different locations in the input feature map. Such similarity may be measured by calculating euclidean distance between features, cosine similarity, and the like. The features are then weighted summed according to the similarity weights to obtain a non-local response for each location. Thus, the network can perform information interaction in a global scope, and thus long-distance dependency relationship is captured.
It should be noted that, in other specific examples of the present application, the fused semantic electric signal waveform shallow feature map may be further subjected to receptive field amplification by other manners to obtain a global semantic electric signal waveform feature map as the electric signal waveform feature, for example: first, the scale of receptive field amplification, i.e., the global context range that is desired to be captured in the feature map, needs to be determined. This scale may be selected based on the requirements of the task and the characteristics of the data, e.g., a larger receptive field may be selected to capture broader context information; the receptive field amplification module is a component for receptive field amplification of the feature map. It may take different forms such as hole convolution, pyramid pooling, etc. These methods can expand receptive field range by adjusting the void fraction of the convolution kernel or the scale of the pooling operation; and inputting the fused semantic electric signal waveform shallow feature map into a receptive field amplification module to perform receptive field amplification operation. The specific manner of operation depends on the receptive field amplification module selected, e.g., when using hole convolution, the extent of receptive fields can be varied by adjusting the hole rate of the convolution kernel; and obtaining a global semantic electric signal waveform characteristic diagram after receptive field amplification operation. This feature map will contain broader context information, enabling better capture of global semantic features.
It should be noted that, in other specific examples of the present application, the waveform feature of the amplified electrical signal may be extracted by other manners to obtain waveform features of the electrical signal, for example: and selecting a proper waveform characteristic extraction method according to the type of the waveform characteristic required to be extracted. Common waveform characteristics include peak, peak-to-peak, mean, standard deviation, rise time, fall time, period, etc.; and sampling the amplified electric signals by using a sampling device or a data acquisition card. The sampling frequency should be selected according to the highest frequency component of the signal to ensure that the sampling is of sufficient detail and accuracy; the sampled analog signal is converted into a digital signal. This may be done by an analog-to-digital converter (ADC) converting the analog signal into a digital representation; the digitized signal is computed using the selected waveform feature extraction method to obtain the desired waveform feature. For example, peaks, averages, standard deviations, etc. of the signals may be calculated; and carrying out data analysis and further application according to the calculation result of the waveform characteristics. Signal classification, fault detection, pattern recognition, etc. can be performed according to specific requirements.
In particular, the nox gas concentration detection module 340 is configured to determine a concentration value of the nox gas based on the waveform characteristics of the electrical signal. In particular, in one specific example of the application, the global semantic electrical signal waveform profile is passed through a decoder to obtain decoded values representing concentration values of the nox gas. The method is characterized in that the global depth fusion semantic feature information of the amplified electric signal is utilized to carry out decoding regression, so that the concentration value of the nitrogen oxide gas is detected, the influence of noise on the weak electric signal can be reduced, and the anti-interference capability of the sensor is enhanced. Specifically, the decoder is used to perform decoding regression on the global semantic electric signal waveform characteristic diagram in the following formula to obtain a decoded value for representing the concentration value of the nitrogen oxide gas; wherein, the formula is:wherein->Representing the global semantic electric signal waveform characteristic diagram, < >>Is the decoded value,/->Is a weight matrix, < >>Representing matrix multiplication.
Notably, decoding regression is a task in machine learning or deep learning that aims to predict a continuous numerical target variable from input data. In decoding regression, the training goal of the model is to minimize the difference between the predicted and actual values, which is typically measured using a loss function. Common loss functions include mean square error, mean absolute error = etc., which can measure the distance or difference between the predicted value and the true value. The model of the decoding regression task may be a variety including linear regression, decision tree regression, support vector regression, neural networks, and the like. Deep learning models perform well in decoding regression tasks because they can learn more complex feature representations and can be trained end-to-end through deep neural networks.
It should be noted that, in other specific examples of the present application, the concentration value of the nitrogen oxide gas may also be determined based on the waveform characteristics of the electrical signal in other manners, for example: collecting electric signal waveform characteristic data related to the concentration of the nitrogen oxide gas; preprocessing the acquired data to ensure the quality and consistency of the data. This may include removing noise, processing missing values, data normalization, etc., to prepare the data for subsequent modeling and prediction; and extracting the characteristic related to the concentration of the nitrogen oxide gas from the waveform characteristic data of the electric signals. This may be achieved by various feature extraction techniques, such as time domain feature extraction, frequency domain feature extraction, wavelet transformation, and the like. The aim is to extract effective features capable of capturing nitrogen oxide concentration information; and using the extracted characteristics as input, taking the concentration of the nitrogen oxide gas as a target variable, and establishing a regression model. Different regression algorithms may be selected, such as linear regression, support vector regression, decision tree regression, neural networks, etc. The selection of the model is based on the characteristics of the data and the requirements of the task; the regression model is trained using the labeled dataset and the performance of the model is assessed using an assessment index (e.g., mean square error, mean absolute error, etc.). If the performance of the model does not meet the requirements, parameter tuning or other models can be tried; and applying a trained regression model to the new electric signal waveform characteristic data to predict the concentration value of the nitrogen oxide gas. The model will generate corresponding concentration prediction results according to the input electric signal waveform characteristics.
It should be appreciated that training of the first convolutional neural network model-based waveform shallow feature extractor, the second convolutional neural network model-based waveform deep feature extractor, the joint semantic propagation module, the non-local neural network model-based feature receptive field amplifier, and the decoder is required prior to the inference using the neural network models described above. That is, the micro-chemiluminescence nox gas sensor 300 based on the internet of things according to the present application further includes a training stage 400 for training the waveform shallow feature extractor based on the first convolutional neural network model, the waveform deep feature extractor based on the second convolutional neural network model, the joint semantic propagation module, the feature receptive field amplifier based on the non-local neural network model, and the decoder.
Fig. 3 is a flowchart of a training phase of a micro-chemiluminescent nitrogen oxide gas sensor based on the internet of things according to an embodiment of the application. As shown in fig. 3, a micro-chemiluminescent method nitrogen oxide gas sensor 300 based on the internet of things according to an embodiment of the present application comprises: training phase 400, comprising: a training data acquisition unit 410 for acquiring training data including a training electrical signal acquired by the optical detection system and a true value of the concentration value of the nitrogen oxide gas; the training signal amplifying unit 420 is configured to amplify the training electrical signal to obtain a training amplified electrical signal; a training waveform shallow feature extraction unit 430, configured to pass the training amplified electrical signal through the waveform shallow feature extractor based on the first convolutional neural network model to obtain a training electrical signal waveform shallow feature map; a training waveform deep feature extraction unit 440, configured to pass the training electric signal waveform shallow feature map through the waveform deep feature extractor based on the second convolutional neural network model to obtain a training electric signal waveform deep feature map; a training depth semantic feature fusion unit 450, configured to use the joint semantic propagation module to fuse the training electrical signal waveform deep feature map and the training electrical signal waveform shallow feature map to obtain a training fusion semantic electrical signal waveform shallow feature map; the training receptive field amplifying unit 460 is configured to pass the training fusion semantic electric signal waveform shallow feature map through the feature receptive field amplifier based on the non-local neural network model to obtain a training global semantic electric signal waveform feature map; the feature activating unit 470 is configured to perform feature rank expression semantic information homogenization activation on the global semantic electric signal waveform feature vector obtained after the training global semantic electric signal waveform feature map is developed, so as to obtain a training global semantic electric signal waveform feature map after activation; a decoding loss unit 480, configured to pass the activated training global semantic electric signal waveform feature map through the decoder to obtain a decoding loss function value; a model training unit 490 for training the waveform shallow feature extractor based on the first convolutional neural network model, the waveform deep feature extractor based on the second convolutional neural network model, the joint semantic propagation module, the feature receptive field amplifier based on the non-local neural network model, and the decoder based on the decoding loss function value and by back propagation of gradient descent.
In particular, in the technical solution of the present application, when the electrical signal waveform deep feature map and the electrical signal waveform shallow feature map are fused by using a joint semantic propagation module to obtain the fused semantic electrical signal waveform shallow feature map, the shallow image semantic features expressed by the electrical signal waveform shallow feature map are weighted by taking the image semantic feature representation of the feature matrix as a unit based on the deep image semantic feature global representation of the electrical signal waveform deep feature map, and when the fused semantic electrical signal waveform shallow feature map passes through the feature feeling wild amplifier based on the non-local neural network model, the global image semantic feature extraction is performed on the image semantic feature representation of the feature matrix, so that the global semantic electrical signal waveform feature map passes through the decoderIn the case of line decoding regression, scale heuristic regression probability mapping is performed based on each feature matrix of the global semantic electric signal waveform feature map, but the training efficiency of the decoder is reduced due to the fact that each feature matrix has a shallow-deep and local-global mixed image semantic feature distribution representation. Based on this, when the applicant decodes the global semantic electric signal waveform feature map through the decoder, semantic information uniformity activation of feature rank expression is performed on the global semantic electric signal waveform feature vector obtained after the global semantic electric signal waveform feature map is developed, specifically expressed as: Wherein (1)>Is the global semantic electric signal waveform feature vector obtained after the training global semantic electric signal waveform feature map is unfolded, < ->Is the +.f of the global semantic electric signal waveform feature vector obtained after the training global semantic electric signal waveform feature map is unfolded>Personal characteristic value->Representing the two norms of the global semantic electric signal waveform feature vector obtained after the training global semantic electric signal waveform feature map is unfolded,/for the global semantic electric signal waveform feature vector>Is a logarithmic function value based on 2, and +.>Is a weight superparameter,/->Is the waveform characteristic diagram display of the post-activation training global semantic electric signalAnd (5) training the waveform feature vector of the global semantic electric signal after activation. Here, the global semantic electric signal waveform feature vector +.>Feature distribution mapping of the feature distribution in the high-dimensional feature space to the decoding regression space can present different mapping modes on different feature distribution levels based on the semantic features of the mixed image, so that the optimal efficiency cannot be obtained based on a scale heuristic mapping strategy, and therefore, rank expression semantic information based on feature vector norms is uniform instead of scale for feature matching, similar feature rank expressions can be activated in a similar manner, and the correlation between feature rank expressions with large difference is reduced, so that the problem of feature vector mismatching of the waveform feature vector of the global semantic electric signal is solved >The problem that the probability expression mapping efficiency of the feature distribution under different spatial rank expressions is low is solved, and the training efficiency of the global semantic electric signal waveform feature diagram when decoded by a decoder is improved. Therefore, the influence of noise on weak electric signals can be reduced, the anti-interference capability of the sensor is enhanced, and the high-precision and high-sensitivity measurement of the concentration of the nitrogen oxide gas is realized.
As described above, the micro-chemiluminescent process nitrogen oxide gas sensor 300 based on the internet of things according to the embodiment of the present application may be implemented in various wireless terminals, for example, a server having a micro-chemiluminescent process nitrogen oxide gas detection algorithm based on the internet of things, or the like. In one possible implementation, the micro-chemiluminescent nitrogen oxide gas sensor 300 based on the internet of things according to the embodiments of the present application may be integrated into a wireless terminal as one software module and/or hardware module. For example, the micro-chemiluminescent nitrogen oxide gas sensor 300 based on the internet of things may be a software module in the operating system of the wireless terminal, or may be an application developed for the wireless terminal; of course, the micro-chemiluminescent nitrogen oxide gas sensor 300 based on the internet of things can be one of a plurality of hardware modules of the wireless terminal.
Alternatively, in another example, the internet of things-based micro-chemiluminescent process nox gas sensor 300 and the wireless terminal may also be separate devices, and the internet of things-based micro-chemiluminescent process nox gas sensor 300 may be connected to the wireless terminal through a wired and/or wireless network and transmit interactive information in a agreed data format.
The foregoing description of the embodiments of the present disclosure has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the various embodiments described. The terminology used herein was chosen in order to best explain the principles of the embodiments, the practical application, or the improvement of technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.
Claims (4)
1. A miniature chemiluminescence method nitrogen oxide gas sensor based on the Internet of things is characterized by comprising:
the data signal acquisition module is used for acquiring the electric signals acquired by the optical detection system;
The signal amplifying module is used for amplifying the electric signal to obtain an amplified electric signal;
the electric signal waveform characteristic extraction module is used for extracting waveform characteristics of the amplified electric signal to obtain electric signal waveform characteristics;
the nitrogen oxide gas concentration detection module is used for determining the concentration value of the nitrogen oxide gas based on the waveform characteristics of the electric signals;
wherein, the electric signal waveform characteristic extraction module includes:
the waveform shallow feature extraction unit is used for extracting shallow features of the amplified electric signals through a waveform shallow feature extractor based on the first deep neural network model so as to obtain an electric signal waveform shallow feature map;
the waveform deep feature extraction unit is used for carrying out deep feature extraction on the electric signal waveform shallow feature map through a waveform deep feature extractor based on a second deep neural network model so as to obtain an electric signal waveform deep feature map;
the joint semantic propagation unit is used for fusing the electric signal waveform deep feature map and the electric signal waveform shallow feature map by using a joint semantic propagation module so as to obtain a fused semantic electric signal waveform shallow feature map;
The receptive field amplifying unit is used for carrying out receptive field amplification on the shallow characteristic map of the fused semantic electric signal waveform so as to obtain a global semantic electric signal waveform characteristic map as the electric signal waveform characteristic;
the first depth neural network model is a first convolution neural network model, and the second depth neural network model is a second convolution neural network model;
wherein, the receptive field amplification unit is used for: the shallow feature map of the fused semantic electric signal waveform passes through a feature receptive field amplifier based on a non-local neural network model to obtain the global semantic electric signal waveform feature map;
wherein, the nitrogen oxide gas concentration detection module is used for: and passing the global semantic electric signal waveform characteristic diagram through a decoder to obtain a decoding value, wherein the decoding value is used for representing the concentration value of the nitrogen oxide gas.
2. The internet of things-based micro-chemiluminescent nox gas sensor of claim 1 further comprising a training module for training the first convolutional neural network model-based waveform shallow feature extractor, the second convolutional neural network model-based waveform deep feature extractor, the joint semantic propagation module, the non-local neural network model-based feature receptive field amplifier, and the decoder.
3. The internet of things-based micro-chemiluminescent process of nitrogen oxide gas sensor of claim 2 wherein the training module comprises:
the training data acquisition unit is used for acquiring training data, wherein the training data comprises training electric signals acquired by the optical detection system and a true value of the concentration value of the nitrogen oxide gas;
the training signal amplifying unit is used for amplifying the training electric signal to obtain a training amplified electric signal;
the training waveform shallow feature extraction unit is used for enabling the electric signal after training amplification to pass through the waveform shallow feature extractor based on the first convolutional neural network model so as to obtain a training electric signal waveform shallow feature map;
the training waveform deep feature extraction unit is used for enabling the training electric signal waveform shallow feature map to pass through the waveform deep feature extractor based on the second convolutional neural network model so as to obtain a training electric signal waveform deep feature map;
the training depth semantic feature fusion unit is used for fusing the training electric signal waveform deep feature map and the training electric signal waveform shallow feature map by using the joint semantic propagation module so as to obtain a training fusion semantic electric signal waveform shallow feature map;
The training receptive field amplifying unit is used for enabling the training fusion semantic electric signal waveform shallow feature map to pass through the feature receptive field amplifier based on the non-local neural network model so as to obtain a training global semantic electric signal waveform feature map;
the feature activation unit is used for carrying out feature rank expression semantic information homogenization activation on the global semantic electric signal waveform feature vector obtained after the training global semantic electric signal waveform feature map is unfolded so as to obtain a training global semantic electric signal waveform feature map after activation;
the decoding loss unit is used for enabling the activated training global semantic electric signal waveform characteristic diagram to pass through the decoder so as to obtain a decoding loss function value;
and the model training unit is used for training the waveform shallow-layer feature extractor based on the first convolutional neural network model, the waveform deep-layer feature extractor based on the second convolutional neural network model, the joint semantic propagation module, the feature receptive field amplifier based on the non-local neural network model and the decoder based on the decoding loss function value and through back propagation of gradient descent.
4. The micro-chemiluminescent nitrogen oxide gas sensor of claim 3 wherein the feature activation unit is configured to: carrying out feature rank expression semantic information homogenization activation on the global semantic electric signal waveform feature vector obtained after the training global semantic electric signal waveform feature map is unfolded by using the following optimization formula to obtain the activated training global semantic electric signal waveform feature map;
Wherein, the optimization formula is:
wherein V is a global semantic electric signal waveform feature vector obtained after the training global semantic electric signal waveform feature map is unfolded, V i Is the ith eigenvalue of the waveform eigenvector of the global semantic electric signal obtained after the waveform eigenvector of the global semantic electric signal is unfolded 2 Representing the two norms of the global semantic electric signal waveform feature vector obtained after the training global semantic electric signal waveform feature map is unfolded, log is a logarithmic function value based on 2, and alpha is a weight super-parameter, v' i The feature vector of the waveform of the activated training global semantic electric signal is obtained after the waveform feature map of the activated training global semantic electric signal is unfolded.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311144880.5A CN117147532B (en) | 2023-09-06 | 2023-09-06 | Micro-chemiluminescence method nitrogen oxide gas sensor based on Internet of things |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311144880.5A CN117147532B (en) | 2023-09-06 | 2023-09-06 | Micro-chemiluminescence method nitrogen oxide gas sensor based on Internet of things |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117147532A CN117147532A (en) | 2023-12-01 |
| CN117147532B true CN117147532B (en) | 2024-03-01 |
Family
ID=88898477
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311144880.5A Active CN117147532B (en) | 2023-09-06 | 2023-09-06 | Micro-chemiluminescence method nitrogen oxide gas sensor based on Internet of things |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117147532B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118054111B (en) * | 2024-02-27 | 2024-09-10 | 武汉怡特环保科技有限公司 | Lithium battery pack safety management method and device, storage medium and electronic equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| SU805145A1 (en) * | 1979-04-26 | 1981-02-15 | Войсковая часть 90720 | Device for chemi-luminescent analysis |
| CN1109167A (en) * | 1993-02-22 | 1995-09-27 | 可口可乐公司 | A method and system for sampling and determining the presence of compoands |
| JP2002148192A (en) * | 2000-11-07 | 2002-05-22 | Shimadzu Corp | Chemiluminescent nitrogen oxide concentration meter |
| CN104316493A (en) * | 2014-11-04 | 2015-01-28 | 中国科学院长春光学精密机械与物理研究所 | Time division multiplexing differential modulation type laser detection method and device for drunk driving |
| CN109283172A (en) * | 2018-11-23 | 2019-01-29 | 大连航创科技有限公司 | A miniaturized NO gas detection device |
| CN211086096U (en) * | 2019-10-18 | 2020-07-24 | 武汉天虹环保产业股份有限公司 | Chemiluminescence method nitrogen oxide analyzer |
-
2023
- 2023-09-06 CN CN202311144880.5A patent/CN117147532B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| SU805145A1 (en) * | 1979-04-26 | 1981-02-15 | Войсковая часть 90720 | Device for chemi-luminescent analysis |
| CN1109167A (en) * | 1993-02-22 | 1995-09-27 | 可口可乐公司 | A method and system for sampling and determining the presence of compoands |
| JP2002148192A (en) * | 2000-11-07 | 2002-05-22 | Shimadzu Corp | Chemiluminescent nitrogen oxide concentration meter |
| CN104316493A (en) * | 2014-11-04 | 2015-01-28 | 中国科学院长春光学精密机械与物理研究所 | Time division multiplexing differential modulation type laser detection method and device for drunk driving |
| CN109283172A (en) * | 2018-11-23 | 2019-01-29 | 大连航创科技有限公司 | A miniaturized NO gas detection device |
| CN211086096U (en) * | 2019-10-18 | 2020-07-24 | 武汉天虹环保产业股份有限公司 | Chemiluminescence method nitrogen oxide analyzer |
Non-Patent Citations (1)
| Title |
|---|
| 基于化学发光检测法的氮氧化物气体分析仪;刘军;冯艳君;刘中军;;仪表技术与传感器;20080315(第03期);第83-84页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117147532A (en) | 2023-12-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111443165B (en) | Odor identification method based on gas sensor and deep learning | |
| CN115791640B (en) | Tumor cell detection equipment and method based on spectroscopic spectrum | |
| CN117104377B (en) | Intelligent management system and method for electric bicycle | |
| CN117147532B (en) | Micro-chemiluminescence method nitrogen oxide gas sensor based on Internet of things | |
| Pan et al. | Identification of complex mixtures for Raman spectroscopy using a novel scheme based on a new multi-label deep neural network | |
| CN116879910B (en) | Laser scanning distance measuring device and method thereof | |
| CN116858789A (en) | Food safety detection system and method thereof | |
| CN117805045A (en) | Aquaculture online monitoring system and method based on image processing technology | |
| CN118169069A (en) | Intelligent water ecological monitoring system and method based on Internet of things | |
| CN119599224B (en) | A complex water quality parameter prediction method and system based on deep learning | |
| CN118839254A (en) | Radiation source individual identification method based on KAN network | |
| CN115049917B (en) | Neural network training method, object detection method and intelligent terminal device | |
| CN111191475A (en) | Passive behavior identification method based on UHF RFID | |
| CN119064712A (en) | A hybrid transmission line fault type identification method based on traveling wave diagnosis and machine vision detection | |
| CN117934963B (en) | Gas sensor drift compensation method | |
| CN110887798B (en) | Nonlinear full-spectrum water turbidity quantitative analysis method based on extreme random tree | |
| CN118673471A (en) | Multispectral water quality prediction method and multispectral water quality prediction system based on self-adaptive random forest | |
| CN118783632A (en) | A method for operating status monitoring based on deep learning and feature screening | |
| CN118916770A (en) | Intelligent individual soldier system based on AI multi-mode | |
| CN118395385A (en) | Ocean environment prediction method and system based on remote sensing and hydrologic sampling data fusion | |
| CN117707116A (en) | Fault diagnosis method, system, computer equipment and medium based on GRU and improved attention mechanism | |
| CN117100191A (en) | Cleaning agent delivery control system and method for automatic dish washing machine | |
| CN117036821A (en) | Single-cell rapid detection and identification method based on laser Raman spectrum | |
| CN110555342B (en) | Image identification method and device and image equipment | |
| CN116539952B (en) | Millimeter wave therapeutic apparatus power intelligent sampling method and system based on Internet of things |
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 |