A cloud collaborative atomized liquid spectrum data intelligent analysis system

Abstract

The application discloses a cloud-end collaborative atomized liquid spectrum data intelligent analysis system and relates to the technical field of atomized spectrum data analysis. The cloud-end collaborative atomized liquid spectrum data intelligent analysis system comprises a spectrum data acquisition module, a cloud-end spectrum processing module and a cloud-end spectrum analysis module. The spectrum data acquisition module is used for acquiring spectrum time series data of a set atomized liquid and uploading the spectrum time series data to a cloud platform for storage. The cloud-end spectrum processing module is used for inputting the spectrum time series data into a pre-trained atomized spectrum analysis model and analyzing spectrum spraying mapping time series feature sets. The cloud-end spectrum analysis module is used for analyzing atomized absorption collaborative characteristic values and atomized dispersion quality characteristic values based on the spectrum spraying mapping time series feature sets of the set atomized liquid. The cloud-end control feedback module is used for jointly spraying control of the set atomized liquid based on the atomized absorption collaborative characteristic values and the atomized dispersion quality characteristic values, so that the cloud-end collaborative atomized liquid spectrum data intelligent analysis system realizes collaborative sensing and dynamic adjustment of atomized liquid characteristics, guarantees the quality of spraying effect, and ensures the accuracy of spraying.

Description

Technical Field

[0001] This invention relates to the field of atomization spectral data analysis technology, specifically to a cloud-based collaborative intelligent analysis system for atomized liquid spectral data. Background Technology

[0002] Atomized liquid refers to a liquid that is dispersed into tiny droplets through a specific device. It has wide applications in industry, agriculture, and medicine. For example, the pesticide atomized liquid produced by precision agricultural equipment such as agricultural drones and intelligent sprayers during operation directly affects the utilization rate, control effect, and environmental impact of pesticides. In recent years, with the rapid development of spectral analysis technology, intelligent analysis of atomized liquids generated during pesticide spraying has been achieved based on spectral data. Spectral data can accurately reflect the physicochemical properties of atomized liquids. Ideally, atomized liquids should have characteristics such as uniform chemical concentration to ensure uniform spraying, good adhesion, and reduced pesticide drift.

[0003] Existing technology, such as the patent application with publication number CN120161113B, discloses an intelligent analysis system and method for atomized liquid components. This method includes: obtaining pyrolysis data through multi-parameter sampling of the atomized liquid; generating a pyrolysis spectrum through temperature gradient analysis; generating a feature library by identifying trace markers; generating a dynamic evolution spectrum by analyzing the component evolution trajectory; generating an atomization feature spectrum through partitioning and delimitation; generating a reaction spectrum through thermodynamic analysis; generating quantitative data by analyzing component diffusion; constructing a sensory response prediction engine; generating response feature data through intelligent analysis; generating a quality report through evaluation; and generating formulation suggestions and performing precise blending through optimization. This method achieves precise analysis and evaluation of atomized liquid components through multi-dimensional analysis, providing a scientific basis for product optimization and effectively balancing safety and taste experience.

[0004] Based on the above findings, the limitations of existing technologies include at least the following problems: Existing technologies lack comprehensive feedback on changes in droplet characteristics during atomized liquid spraying. Atomized liquid is affected by various factors during spraying, causing its optical properties to change accordingly. However, existing technologies struggle to respond to these dynamic changes. The lack of continuous perception of droplet states during spraying makes it difficult to precisely adjust based on droplet absorption and scattering characteristics. This makes it difficult to effectively adapt to uneven droplet distribution and unstable optical properties during atomized liquid spraying, potentially leading to excessively large or small droplet sizes and low spraying efficiency. This, in turn, affects spraying quality and precision, thus limiting the application of atomized liquid spraying in high-precision, dynamically adjustable scenarios. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a cloud-based collaborative intelligent analysis system for atomized liquid spectral data. This system solves the problem that existing technologies lack real-time perception and adjustment of droplet characteristic changes during atomized liquid spraying, resulting in insufficient spraying quality and precision.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a cloud-based collaborative intelligent analysis system for spectral data of atomized liquids, comprising: a spectral data acquisition module for acquiring the spectral time-series data of a set atomized liquid, the spectral time-series data including absorption spectrum time-series data and scattering spectrum time-series data, and uploading it to a cloud platform for storage; a cloud-based spectral processing module for inputting the spectral time-series data of the set atomized liquid stored in the cloud platform into a pre-trained atomization spectral analysis model to analyze the spectral spraying mapping time-series feature set of the set atomized liquid, including the light absorption spraying mapping time-series feature set and the scattering spraying mapping time-series feature set; a cloud-based spectral analysis module for analyzing the atomization absorption synergistic feature value and atomization dispersion quality feature value of the set atomized liquid based on the spectral spraying mapping time-series feature set of the set atomized liquid; and a cloud-based control feedback module for performing joint spraying control of the set atomized liquid based on the atomization absorption synergistic feature value and the atomization dispersion quality feature value.

[0007] Furthermore, the specific steps for uploading the spectral time-series data of the atomizing liquid to the cloud platform for storage are as follows: obtain the network status data of the device equipped with the atomizing liquid at each time point, and analyze its network performance feature set at the corresponding time point, including network connection performance feature value and network load interference feature value; based on the network performance feature set of the device equipped with the atomizing liquid at each time point, analyze its network adaptation feature value at the corresponding time point, and perform upload and storage processing.

[0008] Furthermore, the network status data includes communication strength values, network traffic values, network bandwidth values, signal interference values, and spectrum utilization values. The specific steps for analyzing the network performance characteristic set of the device carrying the atomizing liquid at each time point are as follows: Based on the communication strength values, network bandwidth values, and spectrum utilization values ​​of the device carrying the atomizing liquid at each time point, analyze the network connection performance characteristic values ​​of the corresponding time points; based on the network traffic values ​​and signal interference values ​​of the device carrying the atomizing liquid at each time point, analyze the network load interference characteristic values ​​of the corresponding time points.

[0009] Furthermore, the time-series data of the absorption spectrum includes the wavelength value of each absorption data point at each time point, as well as the corresponding absorptivity and transmittance values; the time-series data of the scattering spectrum includes the wavelength value of each scattering data point at each scattering angle at each time point, as well as the corresponding scattering intensity value; the time-series feature set of the light absorption spraying mapping includes the absorption-transmission balance feature value, the light absorption response intensity feature value, and the light absorption symmetry feature value at each time point; and the scattering spraying mapping feature set includes the scattering concentration feature value, the particle size distribution feature value, and the scattering distribution covariance feature value at each time point.

[0010] Furthermore, the atomization spectral analysis model includes an absorption spectral analysis subnetwork and a scattering spectral analysis subnetwork. The specific steps for analyzing the spectral temporal feature set of the set atomized liquid are as follows: In the absorption spectral analysis subnetwork of the atomization spectral analysis model, the temporal data of the absorption spectrum of the set atomized liquid is received, and the absorption temporal feature is extracted and analyzed to obtain the light absorption spraying mapping temporal feature set of the set atomized liquid; In the scattering spectral analysis subnetwork of the atomization spectral analysis model, the temporal data of the scattering spectrum of the set atomized liquid is received, and the scattering temporal feature is extracted and analyzed to obtain the scattering spraying mapping temporal feature set of the set atomized liquid.

[0011] Furthermore, the absorption spectral analysis sub-network includes an absorption input layer, an absorption extraction layer, and an absorption output layer. The specific steps for analyzing and setting the time-series feature set of the light absorption spraying mapping of the atomized liquid are as follows:

[0012] In the absorption input layer of the absorption spectral analysis subnetwork, the time-series data of the absorption spectrum of the set atomized liquid is received and normalized. In the absorption extraction layer of the absorption spectral analysis subnetwork, the light absorption reaction time-series feature vector of the set atomized liquid is extracted based on the normalized absorption spectrum time-series data of the set atomized liquid. In the absorption output layer of the absorption spectral analysis subnetwork, the light absorption spraying mapping time-series feature set of the set atomized liquid is output based on the light absorption reaction time-series feature vector of the set atomized liquid.

[0013] Furthermore, the scattering spectrum analysis sub-network includes a scattering input layer, a scattering spray feature extraction layer, and a scattering output layer. The specific steps for analyzing the scattering spray mapping time-series feature set of the set atomized liquid are as follows: In the scattering input layer of the scattering spectrum analysis sub-network, the time-series data of the scattering spectrum of the set atomized liquid is received and normalized; in the scattering spray feature extraction layer of the scattering spectrum analysis sub-network, the scattering response time-series feature vector of the set atomized liquid is extracted based on the normalized scattering spectrum time-series data of the set atomized liquid; in the scattering output layer of the scattering spectrum analysis sub-network, the scattering spray mapping time-series feature set of the set atomized liquid is output based on the scattering response time-series feature vector of the set atomized liquid.

[0014] Furthermore, the specific steps for analyzing the synergistic eigenvalues ​​of the atomization absorption of the set atomizing liquid are as follows: Based on the absorption-transmission balance eigenvalues, light absorption response intensity eigenvalues, and light absorption symmetry eigenvalues ​​of the set atomizing liquid at each time point, analyze the initial synergistic eigenvalues ​​of the atomization absorption at the corresponding time points; perform a moving average processing on the initial synergistic eigenvalues ​​of the set atomizing liquid at each time point to obtain the synergistic eigenvalues ​​of the set atomizing liquid.

[0015] Further, the specific steps for analyzing the atomization dispersion quality characteristic values ​​of the set atomizing liquid are as follows: Based on the scattering concentration characteristic value, particle size distribution characteristic value, and scattering distribution covariance characteristic value of the set atomizing liquid at each time point, analyze the initial atomization dispersion quality characteristic value at the corresponding time point; perform a moving average processing on the initial atomization dispersion quality characteristic value of the set atomizing liquid at each time point to obtain the atomization dispersion quality characteristic value of the set atomizing liquid.

[0016] Furthermore, the specific steps for joint spraying control of the set atomizing liquid based on the synergistic characteristic value of atomization absorption and the characteristic value of atomization dispersion are as follows: normalize the synergistic characteristic value of atomization absorption and the characteristic value of atomization dispersion of the set atomizing liquid; compare the normalized synergistic characteristic value of atomization absorption and the characteristic value of atomization dispersion of the set atomizing liquid with several preset atomization spraying adjustment intervals for judgment and analysis; select preset spraying control measures based on the judgment and analysis results.

[0017] The present invention has the following beneficial effects:

[0018] (1) The cloud-based collaborative intelligent analysis system for atomized liquid spectral data acquires and uploads absorption and scattering spectral time series data through the spectral data acquisition module. Using the pre-trained model in the cloud-based spectral processing module, it deeply analyzes the light absorption spraying mapping time series feature set characterizing chemical properties and the scattering spraying mapping time series feature set characterizing physical properties. The cloud-based spectral analysis module then integrates the above time series feature sets to generate atomization absorption collaborative feature values ​​and atomization dispersion quality feature values. The cloud-based control feedback module then makes judgments based on the corresponding feature values ​​and executes precise spraying control, thereby realizing the collaborative perception and dynamic adjustment of atomized liquid properties, thus ensuring the quality of spraying effect and ensuring the accuracy of spraying.

[0019] (2) The cloud-based collaborative intelligent analysis system for atomized liquid spectral data uploads the spectral time series data of the atomized liquid to the cloud platform through the spectral data acquisition module. The cloud platform uses powerful data processing capabilities to store and process the spectral time series data. The cloud spectral analysis module analyzes the stored spectral time series data, extracts the atomization absorption synergy characteristic value and atomization dispersion quality characteristic value, and provides spraying control suggestions based on these characteristics. This enables the system to not only process a large amount of data, but also to optimize and adjust the spraying process, improve the level of spraying intelligence, and enable it to automatically adapt to various environmental changes and application scenarios, thereby improving spraying efficiency and ultimately enhancing the level of spraying intelligence.

[0020] (3) The cloud-based collaborative intelligent analysis system for atomized liquid spectral data normalizes the atomization absorption synergy characteristic value and atomization dispersion quality characteristic value, and then matches the processed dual characteristic values ​​with multiple preset atomization spraying adjustment ranges. The system automatically triggers the corresponding control strategy based on the matching results, thereby taking into account the balance of chemical properties and physical state, and selecting appropriate spraying control measures to achieve precise control of the spraying process, so as to ensure that the best spraying effect can be maintained under different operating environments, thereby significantly improving the pesticide utilization rate.

[0021] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0022] Figure 1 This is a block diagram of a cloud-based collaborative intelligent analysis system for spectral data of atomized liquids according to the present invention.

[0023] Figure 2 The flowchart illustrates the specific steps involved in uploading the spectral time-series data of the atomized liquid to a cloud platform for storage in a cloud-collaborative intelligent analysis system for atomized liquid spectral data according to the present invention.

[0024] Figure 3 This invention provides a flowchart outlining the specific steps involved in analyzing and setting the spectral temporal feature set of an atomizing liquid in a cloud-based collaborative intelligent analysis system for atomizing liquid spectral data.

[0025] Figure 4 This is a schematic diagram of the time-series feature set data of the scattering and spraying mapping of the atomized liquid in a cloud-collaborative intelligent analysis system for atomized liquid spectral data of the present invention. Detailed Implementation

[0026] Please see Figure 1This invention provides a technical solution: a cloud-based collaborative intelligent analysis system for atomized liquid spectral data, comprising: a spectral data acquisition module, used to acquire, within a set period (e.g., spectral time series data) of a set atomized liquid (e.g., atomized liquid generated by a drone spraying pesticides), including absorption spectrum time series data and scattering spectrum time series data, and upload it to a cloud platform for storage; a cloud-based spectral processing module, used to input the spectral time series data of the set atomized liquid stored in the cloud platform into a pre-trained atomization spectral analysis model, and analyze the spectral spraying mapping time series feature set of the set atomized liquid, including absorption spraying mapping time series feature set and scattering spraying mapping time series feature set; a cloud-based spectral analysis module, used to analyze the atomization absorption synergy feature value and atomization dispersion quality feature value of the set atomized liquid based on the spectral spraying mapping time series feature set of the set atomized liquid; and a cloud-based control feedback module, used to perform joint spraying control of the set atomized liquid based on the atomization absorption synergy feature value and atomization dispersion quality feature value.

[0027] The time-series data of the absorption spectrum includes the wavelength value of each absorption data point at each time point, as well as the corresponding absorptivity and transmittance values. The time-series data of the scattering spectrum includes the wavelength value of each scattering data point at each scattering angle at each time point, as well as the corresponding scattering intensity value. The time-series feature set of the light absorption spraying map includes the absorption-transmission balance feature value, the light absorption response intensity feature value, and the light absorption symmetry feature value at each time point. The feature set of the scattering spraying map includes the scattering concentration feature value, the particle size distribution feature value, and the scattering distribution covariance feature value at each time point.

[0028] Furthermore, the pre-training steps for the atomization spectral analysis model are as follows:

[0029] A fogging spectral annotation dataset was obtained, consisting of spectral data from several spraying experiments and droplet information data during the spraying process. This data was annotated by experimental experts based on on-site monitoring records and feedback from the spraying environment. Each sample group includes absorption and scattering spectral time-series data for each moment within a designated area, along with corresponding feature labels, including absorption spraying mapping feature sets and scattering spraying mapping feature sets. Each sample in the dataset has complete annotations, including feature values ​​of the time-series data and target labels. The label values ​​are derived from rigorous spectral analysis and calculation. The fogging spectral annotation dataset is divided into training, validation, and test sets in chronological order, typically with 80% used for training, 10% for validation, and 10% for testing, ensuring the dataset is arranged in chronological order.

[0030] The atomization spectral analysis model is trained. The model includes an absorption spectral analysis subnetwork and a scattering spectral analysis subnetwork. Taking the absorption spectral analysis subnetwork as an example, the absorption spectral time series data at each time point is input into the network and normalized. A deep neural network (such as a convolutional neural network) is used to extract the light absorption response time series feature vector from the normalized data. The network automatically learns the correlation features between the light absorption response and the time series at each time point, captures the light absorption characteristics of the droplets, and converts the light absorption response time series feature vector into absorption-transmission balance feature values, light absorption response intensity feature values, and light absorption symmetry feature values ​​through an activation function (such as sigmoid). The output is a light absorption spraying mapping time series feature set. The goal is to predict the light absorption response characteristics of the droplets by optimizing the model.

[0031] During training, optimization algorithms (such as the Adam optimizer) are used to minimize the loss function. The model's performance is further improved by adjusting hyperparameters such as the learning rate and the number of hidden units in the LSTM layer. The model is evaluated using a validation set to determine the optimal hyperparameter settings, ensuring that the model can effectively extract useful features from unseen data.

[0032] After training is completed, the generalization ability of the model is evaluated using a test set to ensure that it can maintain high prediction accuracy on new data. The model performance is further optimized through model performance evaluation (such as accuracy, precision, recall, etc.). Finally, the trained model parameters are saved for online deployment and real-time evaluation in actual spraying processes.

[0033] Specifically, such as Figure 2 As shown, the specific steps for uploading the spectral time-series data of the atomizing liquid to the cloud platform for storage are as follows: Obtain the network status data of the device carrying the atomizing liquid (communication network with the cloud platform) at each time point, and analyze its network performance characteristic set at the corresponding time point, including network connection performance characteristic values ​​and network load interference characteristic values; Based on the network performance characteristic set of the device carrying the atomizing liquid at each time point, analyze its network adaptation characteristic value at the corresponding time point, that is, perform weighted processing on the network connection performance characteristic value and network load interference characteristic value at each time point, and in this weighted processing, take the reciprocal of the network load interference characteristic value and normalize it, i.e., 1 / (1+network load interference characteristic value), to obtain its network adaptation characteristic value at the corresponding time point, and then upload and store the spectral data at the corresponding time point, specifically as follows:

[0034] The system determines whether the network adaptation feature value of the device carrying the atomizing liquid at each time point is lower than the preset network adaptation feature threshold. If it is lower than the preset network adaptation feature threshold, the spectral data at that time point is compensated (by interpolation methods such as linear interpolation or spline interpolation) or resampled (by upsampling or downsampling), and the processed spectral data is uploaded to the cloud platform for storage, and the timestamp of that time point is recorded. If it is not lower than the preset network adaptation feature threshold, the spectral data at that time point is directly uploaded to the cloud platform (using network protocols such as HTTP / HTTPS or MQTT for transmission), and stored in the spectral database (such as a NoSQL database) of the cloud platform, while the timestamp of that time point is recorded.

[0035] It should be noted that in this implementation example, the weighting process can be cumulative weighting, and the weights of each parameter during the weighting process can be set using particle swarm optimization algorithms, entropy weighting methods, etc. Here, we take the weights of network connectivity performance features and network load interference features obtained through the entropy weighting method as an example:

[0036] Historical data from multiple consecutive time points are collected to form a dataset consisting of a sequence of network connectivity performance feature values ​​and a sequence of network load interference feature values. Assume a total of n valid data points are collected, each containing two evaluation indicators: a network connectivity performance feature value and a preprocessed network load interference feature value. The data from these n time points and the two indicators are arranged into an initial matrix of n rows and two columns. This matrix is ​​then standardized, mapping each indicator value to the interval between zero and one. Next, the proportion of each feature value at different time points is calculated. For each feature value, the ratio of its standardized value at each time point to the sum of its standardized values ​​at all time points is calculated. The weighting is calculated by analyzing the information entropy of each feature value. For each feature value, based on the corresponding weight sequence obtained from the above analysis, its information entropy value is calculated (the information entropy formula can be used). The difference coefficient of each feature value is then analyzed by subtracting the information entropy of the feature value from one. The difference coefficient of each feature value is then normalized to obtain the final weight of the feature value in the weighted calculation. The difference coefficient of a feature value is divided by the sum of the difference coefficients of all feature values ​​(two in this example). The quotient is the weight coefficient of the feature value. Thus, the weights of the network connectivity performance feature value and the network load interference feature value can be obtained respectively, and the sum of the two weights is one.

[0037] Network status data includes communication strength, network traffic, network bandwidth, signal interference, and spectrum utilization. The specific steps for analyzing the network performance characteristic set of the device carrying the atomizing liquid at each time point are as follows: Based on the communication strength, network bandwidth, and spectrum utilization values ​​of the device carrying the atomizing liquid at each time point, analyze the network connection performance characteristic value for that time point. Specifically, standardize the communication strength, network bandwidth, and spectrum utilization values ​​of the device carrying the atomizing liquid at each time point, and then perform weighted processing based on the standardization results to obtain the network connection performance characteristic value (used to characterize the network connection quality between the device and the cloud platform). Based on the network traffic and signal interference values ​​of the device carrying the atomizing liquid at each time point, analyze the network load interference characteristic value for that time point. Specifically, standardize the network traffic and signal interference values ​​of the device carrying the atomizing liquid at each time point, and then perform weighted processing based on the standardization results to obtain the network load interference characteristic value (used to characterize the negative impact of network load and interference levels on transmission performance at that time point).

[0038] The communication strength value is the signal strength of the communication network (such as Wi-Fi, 4G / 5G, etc.) between the device and the cloud platform at that point in time, which can be obtained through the device's built-in wireless signal receiver.

[0039] Network traffic refers to the amount of data transmitted between the device and the cloud platform at a given point in time. It represents the real-time amount of data being transmitted over the network and can be monitored in real time by the device through its built-in flow meter (or flow monitoring sensor).

[0040] Network bandwidth is the maximum data transfer rate available between the device and the cloud platform, which can be obtained using bandwidth testing tools between the device and the cloud platform (such as Speedtest).

[0041] Signal interference value is the degree of interference to wireless signals in the environment where the device is located. It can be obtained by scanning the radio frequency bands in the environment where the device is located through the built-in spectrum analyzer of the device, monitoring the background noise in the frequency band in real time (usually from other wireless devices, electronic devices, etc.). The spectrum analyzer measures the signal strength of each frequency band, including the strength of the effective signal and the strength of the interference signal, and performs comprehensive processing (i.e., SNR) to use it as the signal interference value.

[0042] The spectrum utilization rate is the degree to which the device uses the spectrum bandwidth at a given point in time, i.e., the bandwidth occupancy of the frequency band used by the device. It can be obtained by scanning and analyzing the wireless frequency band used by the device through the built-in spectrum analyzer. Specifically, it is based on the spectrum analyzer measuring the current frequency band bandwidth and the total bandwidth of the frequency band (which can be obtained from the frequency band standards stored in the database), and performing ratio processing, and the result is used as the spectrum utilization rate value.

[0043] In this implementation scheme, the network status between the device and the cloud platform is monitored and analyzed in real time. This allows for intelligent adjustment of the uploading process of the atomized liquid spectral time-series data, ensuring the reliability of data transmission. Secondly, through in-depth analysis of the network status data, network adaptation characteristic values ​​are generated to set the communication network between the atomized liquid device and the cloud platform, and dynamic judgment is performed. When poor network conditions are detected, the system will proactively compensate or optimize the spectral data to ensure data integrity and continuity. This effectively avoids data loss or incomplete transmission problems that may be caused by network fluctuations, thus ensuring that the data received from the cloud always has high availability. Finally, the system selects the optimal transmission strategy based on the network adaptation characteristic values, ensuring both timely uploading of key data and rational utilization of network resources. This enables the entire system to maintain stable and reliable data transmission capabilities under various network environments, thereby ensuring that the atomized liquid spectral data can be successfully uploaded to the cloud platform and provide real-time feedback.

[0044] Specifically, such as Figure 3 As shown, the atomization spectral analysis model includes an absorption spectral analysis subnetwork and a scattering spectral analysis subnetwork. The specific steps for analyzing the spectral temporal feature set of the set atomized liquid are as follows: In the absorption spectral analysis subnetwork of the atomization spectral analysis model, the temporal data of the absorption spectrum of the set atomized liquid is received, and the absorption temporal feature is extracted and analyzed to obtain the light absorption spraying mapping temporal feature set of the set atomized liquid; In the scattering spectral analysis subnetwork of the atomization spectral analysis model, the temporal data of the scattering spectrum of the set atomized liquid is received, and the scattering temporal feature is extracted and analyzed to obtain the scattering spraying mapping temporal feature set of the set atomized liquid.

[0045] The absorption spectral analysis subnetwork comprises an absorption input layer, an absorption extraction layer, and an absorption output layer. The specific steps for analyzing the time-series feature set of the light absorption spray mapping of the set atomized liquid are as follows: In the absorption input layer of the absorption spectral analysis subnetwork, the time-series data of the absorption spectrum of the set atomized liquid is received and normalized, that is, the wavelength value of each absorption data point, along with the corresponding absorptivity and transmittance values, are normalized to between 0 and 1; In the absorption extraction layer of the absorption spectral analysis subnetwork, based on the normalized time-series data of the absorption spectrum of the set atomized liquid… The light absorption response time sequence feature vector of the set atomized liquid is extracted. In the absorption output layer of the absorption spectrum analysis sub-network, based on the light absorption response time sequence feature vector of the set atomized liquid, the light absorption spraying mapping time sequence feature set of the set atomized liquid is output. Specifically, the absorption transmission balance feature, light absorption response intensity feature, and light absorption symmetry feature of each time point in the light absorption response time sequence feature vector are activated by the Sigmoid function to obtain the absorption transmission balance feature value, light absorption response intensity feature value, and light absorption symmetry feature value between 0 and 1.

[0046] The specific steps for extracting the light absorption reaction time sequence feature vector of the atomized liquid are as follows: Read the absorptivity and transmittance values ​​of each absorption data point at each time point, and perform ratio processing to obtain the absorption transmittance of the corresponding absorption data point. Extract the maximum, minimum, mean, and standard deviation of the absorption transmittance. Perform difference processing on the maximum and minimum absorption transmittance values ​​to obtain the range of absorption transmittance. Ratio the standard deviation with the mean absorption transmittance, and weight the result with the range of absorption transmittance. Take the reciprocal of this weighted result and normalize it to 1 / (1+this weighted result) to extract the absorption transmittance balance feature at each time point. This feature is used to characterize the optical balance of the pesticide liquid during spraying. The larger this feature, the more balanced the absorption and transmittance characteristics of the liquid, the more stable the optical properties of the liquid, and the better the balance of absorbed and transmitted light during spraying, thus contributing to uniform coverage of the spraying area.

[0047] Read the wavelength and absorbance values ​​of each absorption data point at each time point. Extract the absorbance change rate between adjacent data points sequentially, such as the absolute value of the difference between the absorbance values ​​of the first and second data points divided by the absorbance value of the second data point. Calculate the mean and standard deviation of the absorbance change rate, and use this to set an absorbance change threshold, i.e., mean absorbance change rate + 2 × standard deviation of absorbance change rate. Iterate through the absorbance change rates of all data points. Divide the first data point among two adjacent data points whose absorbance change rate exceeds the set threshold into intervals, and then... Two data points are used as the starting point of a new interval until the rate of change of absorbance exceeds the set absorbance change threshold, thus obtaining several band intervals. For the wavelength value and absorbance value of each data point in each band interval, the least squares method is used to fit a straight line to extract the slope value of the fitted straight line for each band interval. The slope value is then weighted to extract the absorbance response intensity feature at each time point, which is used to characterize the absorption sensitivity of the pesticide liquid at different wavelengths. The larger the feature, the higher the sensitivity of the liquid to different wavelengths, thus assessing the more significant the liquid's response to external changes during spraying.

[0048] Read the absorptivity value of each absorption data point at each time point and perform peak detection processing (such as using local maximum detection or Gaussian fitting). Taking local maximum detection as an example, for any data point's absorptivity value, compare it with the absorptivity values ​​of its two adjacent data points. If the absorptivity value of the data point is higher than the absorptivity values ​​of its two adjacent data points, then mark the data point as a candidate peak. Count the maximum absorptivity value among all candidate peaks, and mark the candidate peak corresponding to the maximum absorptivity value as the actual peak. Take half of the absorptivity value corresponding to the actual peak, i.e., the half-maximum absorptivity value. Then, count the data points on the left and right sides of the actual peak corresponding to the half-maximum absorptivity value (it should be noted that if no corresponding data point is found, linear interpolation is used to obtain the wavelength values ​​of the left and right data points). For data points, the wavelength values ​​of the actual peak point and the left and right data points are compared (absolute value is taken) to obtain the left and right half widths. The ratio is then calculated as |left half width - right half width| / (left half width + right half width) to obtain the width ratio. The absorbance values ​​of all data points to the left of the actual peak point are multiplied by the wavelength interval to obtain the left area. Similarly, the right area is obtained and its ratio is calculated as |left area - right area| / (left area + right area) to obtain the area ratio. This ratio is then weighted with the width ratio, and the reciprocal of the weighted result is normalized to 1 / (1 + the weighted result) to extract the absorbance symmetry feature at each time point. This feature is used to characterize the uniformity of pesticide liquid during spraying. The larger the feature, the better the uniformity of liquid spraying, reflecting a more uniform distribution of spray dosage and concentration.

[0049] The absorption and transmission balance characteristics, light absorption response intensity characteristics, and light absorption symmetry characteristics at each time point are concatenated into a light absorption spraying response characteristic vector at each time point, which is the light absorption response time sequence characteristic vector.

[0050] The scattering spectrum analysis subnetwork includes a scattering input layer, a scattering spray feature extraction layer, and a scattering output layer. The specific steps for analyzing the scattering spray mapping time-series feature set of the set atomized liquid are as follows: In the scattering input layer of the scattering spectrum analysis subnetwork, the time-series data of the scattering spectrum of the set atomized liquid is received and normalized; in the scattering spray feature extraction layer of the scattering spectrum analysis subnetwork, the scattering response time-series feature vector of the set atomized liquid is extracted based on the normalized scattering spectrum time-series data of the set atomized liquid.

[0051] In the scattering output layer of the scattering spectrum analysis subnetwork, based on the set scattering response time-series feature vector of the atomized liquid, the set scattering spraying mapping time-series feature set of the atomized liquid is output. Specifically, the scattering concentrated feature, particle size distribution feature, and scattering distribution covariant feature at each time point in the scattering response time-series feature vector are activated by the Sigmoid function to obtain the scattering concentrated feature value, particle size distribution feature value, and scattering distribution covariant feature value between 0 and 1.

[0052] The specific steps for extracting the temporal feature vector of the scattering response of the atomizing liquid are as follows: Read the wavelength value and corresponding scattering intensity value of each scattering data point at each scattering angle at each time point, and perform the following analysis for any time point:

[0053] The scattering angle is divided into several angular intervals (e.g., 0°-20°, 20°-40°, 40°-60°, and so on). The scattering intensity and value for each angular interval are calculated (i.e., the sum of the scattering intensity values ​​of all scattering data points within that angular interval). The maximum and total scattering intensities are extracted from this data, and ratios are applied to extract the scattering aggregation value. Based on the scattering intensity and value for each angular interval, the mean and standard deviation of the scattering intensity are extracted. Using the skewness formula, the scattering distribution at each time point is extracted. The skewness value is weighted with the scattering aggregation value to extract the scattering concentration feature, which is used to characterize the concentration and symmetry of the scattered energy distribution of atomized droplets in space. The larger the value, the more concentrated the scattered light energy is in certain specific forward or backward angle ranges, and the greater the skewness of the angle distribution shape. This means that the droplet size distribution is uneven, large droplets exist, or agglomeration occurs. Conversely, the smaller the feature value, the more uniform and symmetrical the scattered light energy distribution in space, reflecting that the droplet size is uniform, the atomization state is stable, and the dispersion is good.

[0054] The wavelength is divided into several intervals (e.g., 200nm-400nm, 400nm-600nm, 600nm-800nm, and so on). Each wavelength interval represents a different characteristic of droplet size distribution. Within each wavelength interval, the scattering intensity values ​​of all scattering data points are statistically analyzed, and the sum of scattering intensities for that wavelength interval is analyzed. For the sums of scattering intensities of adjacent wavelength intervals, a comprehensive analysis is performed, i.e., |difference in the sum of scattering intensities of adjacent wavelength intervals| / |difference in the upper limit of adjacent wavelength intervals|, resulting in several sets of adjacent... The scattering intensity variation values ​​within the wavelength range are used to extract the mean scattering intensity variation. Simultaneously, the sum of scattering intensities in each group of adjacent wavelength ranges is processed by ratio, and the variance of the result is calculated. This result is then standardized with the mean scattering intensity variation. A weighted average is performed based on the standardized result to extract the particle size distribution characteristics at each time point. This characteristic value is used to characterize the uniformity of the atomized droplet particle size distribution. The smaller this characteristic value, the more concentrated the scattering intensity appears in certain specific wavelength ranges, which means that the droplet particle size distribution is more concentrated and uniform, closer to a monodisperse state, and the higher the atomization quality.

[0055] For each scattering data point at each scattering angle and its corresponding scattering intensity value, the correlation between wavelength and scattering intensity at a certain angle is analyzed based on the Pearson correlation coefficient. The mean, standard deviation, and extreme values ​​of the wavelength-scattering intensity correlation are extracted and processed. A weighted average is then applied to extract the covariant similarity feature of the scattering distribution at each time point. This feature is used to characterize the similarity of the spectral shape at different scattering angles. The larger the feature value, the more similar the shapes of the wavelength and scattering intensity spectral curves measured at all scattering angles are, reflecting the stability of the droplet system's physicochemical properties and the uniformity of its particle size distribution. The concentrated scattering feature, particle size distribution feature, and covariant scattering distribution feature at each time point are concatenated into a scattering spray response feature vector at each time point, i.e., a scattering response time-series feature vector.

[0056] In this implementation scheme, the absorption spectral analysis subnetwork and the scattering spectral analysis subnetwork are used to perform in-depth analysis of the time-series data of the light absorption spectrum and the scattering spectrum of the set atomizing liquid, respectively. This allows for the comprehensive extraction of characteristic features related to the set atomizing liquid. For example, the absorption spectral analysis subnetwork can accurately capture the optical balance and distribution uniformity of the liquid, while the scattering spectral analysis subnetwork effectively analyzes the spatial distribution concentration and particle size uniformity of the droplets. This ensures that the system can fully perceive the atomization state of the atomizing liquid and thus ensures that the droplet characteristics during the spraying process are within a controllable range. This results in a highly efficient and uniform atomization effect, avoiding a decrease in spraying quality and efficiency due to uneven droplet distribution or excessively large particle size, and ensuring that the best spraying effect is maintained under different operating environments.

[0057] Specifically, the steps for analyzing the synergistic atomization absorption characteristic value of the set atomizing liquid are as follows: Based on the absorption-transmission balance characteristic value, light absorption response intensity characteristic value, and light absorption symmetry characteristic value of the set atomizing liquid at each time point, analyze the initial synergistic atomization absorption characteristic value at the corresponding time point (used to characterize the comprehensive quality of the chemical spraying state of the atomizing liquid; the higher the value, the better the chemical spraying state, reflecting the uniform concentration distribution and precise and stable dosage of the pesticide liquid; the lower the value, the worse the chemical spraying state, reflecting uneven concentration or dosage fluctuation); perform a moving average processing on the initial synergistic atomization absorption characteristic value of the set atomizing liquid at each time point (the initial synergistic atomization absorption characteristic value of all time points within the period can be processed based on the moving average formula) to obtain the synergistic atomization absorption characteristic value of the set atomizing liquid.

[0058] The specific formula for calculating the initial synergistic eigenvalue of the atomizing fluid at a certain time point is as follows: ;in, To set the initial synergistic characteristic value of atomization absorption at a certain time point of the atomizing fluid, To set the absorption-transmission equilibrium characteristic value of the atomizing fluid at a certain time point, To set the characteristic value of the light absorption response intensity of the atomizing liquid at a certain time point, To set the absorbance symmetry characteristic value of the atomizing liquid at a certain time point, These are the coordination coefficients stored in the database. These are the variation adjustment coefficients stored in the database. To set the coefficient of variation of the atomizing liquid at a certain time point, it is calculated by dividing the standard deviation of the absorption-transmission equilibrium characteristic value, the absorbance response intensity characteristic value, and the absorbance symmetry characteristic value by their average value. In this embodiment, the co-regulation coefficient stored in the database... Variable adjustment coefficient The values ​​are 1.500 and 2.000 respectively.

[0059] The specific steps for analyzing the atomization dispersion quality characteristic values ​​of a set atomizing liquid are as follows: Based on the scattering concentration characteristic value, particle size distribution characteristic value, and scattering distribution covariance characteristic value of the set atomizing liquid at each time point, analyze the initial atomization dispersion quality characteristic value at the corresponding time point (used to characterize the comprehensive quality of the physical dispersion state of the atomizing liquid; the higher the value, the better the physical dispersion state, reflecting uniform droplet size distribution and symmetrical spatial distribution; the lower the value, the worse the physical dispersion state, reflecting uneven droplet size or agglomeration); perform a moving average processing on the initial atomization dispersion quality characteristic value of the set atomizing liquid at each time point (the initial atomization dispersion quality characteristic value of all time points within the period can be processed based on the moving average formula) to obtain the atomization dispersion quality characteristic value of the set atomizing liquid.

[0060] The specific formula for calculating the initial atomization dispersion quality characteristic value of the atomizing liquid at a certain time point is as follows: ;in, To set the initial atomization dispersion quality characteristic value of the atomizing liquid at a certain time point, To set the concentrated eigenvalue of the scattering at a certain time point of the atomizing liquid, The scattering concentration adjustment coefficients are stored in the database. To set the particle size distribution characteristic value of the atomizing liquid at a certain time point, The granularity distribution adjustment coefficients are stored in the database. To define the covariant eigenvalue of the scattering distribution of the atomizing liquid at a certain time point, These are the covariant adjustment coefficients for the scattering distribution stored in the database. Furthermore, in this embodiment, the scattering concentration adjustment coefficients stored in the database Particle size distribution adjustment coefficient Covariance adjustment coefficient of scattering distribution The values ​​were 0.386, 0.325, and 0.289, respectively.

[0061] The following is a specific implementation example for calculating the initial atomization dispersion quality characteristic values ​​of a set atomizing liquid at a certain time point. The available data includes the scattering concentration characteristic values, particle size distribution characteristic values, and scattering distribution covariance characteristic values ​​for five randomly selected time points of the set atomizing liquid, as detailed in Table 1 and... Figure 4 As shown:

[0062] Table 1. Example of time-series feature set data for setting the scattering and spraying mapping of atomized liquid.

[0063] Scattering concentrated eigenvalues Particle size distribution characteristic value Covariant eigenvalues ​​of scattering distribution Time point 1 0.265 0.314 0.736 Time point 2 0.243 0.296 0.695 Time point 3 0.317 0.348 0.824 Time point 4 0.386 0.357 0.796 Time point 5 0.253 0.327 0.758

[0064] Scattering lumbar adjustment coefficients stored in the database The value is: 0.386;

[0065] Granularity distribution adjustment coefficients stored in the database The value is: 0.325;

[0066] Covariant adjustment coefficients of scattering distribution stored in the database The value is: 0.289;

[0067] Substituting the data from Table 1 and the coefficients mentioned above into the specific formula for calculating the initial atomization dispersion quality characteristic value of the atomizing liquid at a certain time point, we obtain:

[0068] The initial atomization dispersion quality characteristic value of the atomizing liquid at the first time point is set as follows: 0.386×exp(-0.265)+0.325×(1 / (1+0.314))+0.289×√0.736≈0.792;

[0069] The initial atomization dispersion quality characteristic value of the atomizing liquid at the second time point is set as 0.386×exp(-0.243)+0.325×(1 / (1+0.296))+0.289×√0.695≈0.794;

[0070] The initial atomization dispersion quality characteristic value of the atomizing liquid at the third time point is set as 0.386×exp(-0.317)+0.325×(1 / (1+0.348))+0.289×√0.824≈0.785;

[0071] The initial atomization dispersion quality characteristic value of the atomizing liquid at the fourth time point is set as 0.386×exp(-0.386)+0.325×(1 / (1+0.357))+0.289×√0.796≈0.759;

[0072] The initial atomization dispersion quality characteristic value of the atomizing liquid at the fifth time point is set as 0.386×exp(-0.253)+0.325×(1 / (1+0.327))+0.289×√0.758≈0.796.

[0073] In this implementation scheme, by comprehensively analyzing the time-series feature sets of light-absorbing spraying and light-scattering spraying, the chemical and physical spraying state of the atomized liquid can be fully evaluated. Secondly, the features in the time-series feature set of light-absorbing spraying can help determine the optical balance and uniformity of the liquid during the spraying process. For example, by extracting features such as light absorption response intensity and symmetry, changes in the optical properties of the droplets can be monitored to ensure that the light absorption and transmission characteristics are maintained at their optimal state during the spraying process. The features in the time-series feature set of light-scattering spraying help determine the dispersion state of the droplets. The system can understand whether the droplets are uniformly distributed and whether there are large droplets or agglomeration phenomena. Finally, by comprehensively processing these features, atomization absorption synergy feature values ​​and atomization dispersion quality feature values ​​can be generated, which can be dynamically adjusted throughout the spraying process, thereby significantly improving the spraying accuracy.

[0074] Specifically, the steps for joint spraying control of the set atomizing liquid based on the atomization absorption synergy characteristic value and the atomization dispersion quality characteristic value are as follows: The atomization absorption synergy characteristic value and the atomization dispersion quality characteristic value of the set atomizing liquid are normalized (i.e., the results of the atomization absorption synergy characteristic value and the atomization dispersion quality characteristic value are normalized to between 0 and 1); the normalized atomization absorption synergy characteristic value and the atomization dispersion quality characteristic value of the set atomizing liquid are compared with several preset atomization spraying adjustment intervals for judgment and analysis. Each atomization spraying adjustment interval includes one atomization absorption synergy interval and one atomization dispersion quality interval, and each atomization spraying adjustment interval corresponds to one atomization spraying control measure.

[0075] Based on the judgment and analysis results, preset spraying control measures are selected. These measures are based on the normalized atomization absorption synergistic characteristic value and atomization dispersion quality characteristic value of the set atomizing liquid being within a preset atomization spraying adjustment range. The set atomizing liquid is then subjected to atomization spraying control processing, including but not limited to the following examples:

[0076] Interval Group 1 (High Collaboration, High Quality):

[0077] Synergistic absorption range of atomization: 0.8-1.0 (indicating excellent chemical state, uniform concentration, and stable dosage);

[0078] Atomization dispersion quality range: 0.8-1.0 (indicating excellent physical state, uniform particle size, and symmetrical distribution);

[0079] Corresponding atomization spraying control measures: Maintain all current operating parameters (such as pressure, nozzle orifice diameter, mixing ratio, etc.), and enable energy-saving mode to appropriately reduce the power of auxiliary systems, thereby optimizing energy consumption while maintaining the best spraying effect;

[0080] Interval group 2 (high collaboration, low quality):

[0081] Synergistic absorption range of atomization: 0.7-1.0 (indicating good chemical state);

[0082] Atomization dispersion quality range: 0.0-0.4 (indicating poor physical state, uneven particle size, or agglomeration);

[0083] Corresponding atomization spray control measures: If the chemical composition is uniform but the atomization effect is poor, physical parameters should be adjusted first, such as automatically increasing the spray pressure, checking and cleaning the nozzles or using the backup nozzles, and optimizing the airflow speed of gas-assisted atomization to improve the physical dispersion of droplets.

[0084] Interval group 3 (low collaboration, high quality):

[0085] Synergistic absorption range of atomization: 0.0-0.4 (indicating poor chemical state, uneven concentration, or dose fluctuation);

[0086] Atomization dispersion quality range: 0.7-1.0 (indicating good physical condition);

[0087] Corresponding atomization spraying control measures: If the atomization morphology is good but the chemical composition is uneven, the chemical parameters should be adjusted first, such as automatically adjusting the mixing ratio valve of pesticide concentrate and diluent, increasing the stirring rate of the mixing device, and ensuring the uniformity of chemical concentration of the pesticide solution and the accuracy of the dosage.

[0088] Interval group 4 (medium coordination, medium quality):

[0089] Synergistic absorption range of atomization: 0.4-0.6 (indicating moderate chemical state);

[0090] Atomization dispersion quality range: 0.4-0.6 (indicating moderate physical conditions);

[0091] Corresponding atomization spray control measures: The system is at a medium level and needs comprehensive optimization, such as minor synchronous adjustments to pressure and mixing parameters;

[0092] Interval group 5 (low collaboration, low quality):

[0093] Synergistic absorption range of atomization: 0.0-0.3 (indicating very poor chemical state);

[0094] Atomization dispersion quality range: 0.0-0.3 (indicating very poor physical state);

[0095] Corresponding atomization spraying control measures: If the system deteriorates completely, immediately implement a comprehensive intervention strategy, such as simultaneously and significantly adjusting the mixing ratio, increasing the stirring rate, increasing the spraying pressure, performing nozzle cleaning cycles or switching to backup pipelines, and triggering audible and visual alarms to prompt the operator to perform manual intervention and system checks.

[0096] In this implementation scheme, spraying quality can be significantly improved by combining the atomization absorption synergistic characteristic value and the atomization dispersion quality characteristic value for spraying control. Furthermore, by dividing the atomization absorption synergistic characteristic value and the atomization dispersion quality characteristic value into different ranges, the system can automatically select appropriate control measures according to different spraying conditions, thereby achieving precise control of the atomized liquid state and ensuring that the best atomization effect is maintained in different spraying scenarios. For example, if the liquid has a good chemical state but poor physical distribution, the system will automatically increase the pressure or clean the nozzle to improve the atomization effect, thereby enabling it to respond to changes in droplet state and make flexible adjustments to ensure stability during the spraying process.

[0097] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the invention.

[0098] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A cloud-based collaborative intelligent analysis system for spectral data of atomized liquids, characterized in that, include: The spectral data acquisition module is used to acquire the spectral time-series data of the set atomizing liquid. The spectral time-series data includes absorption spectrum time-series data and scattering spectrum time-series data, which are then uploaded to a cloud platform for storage. Specifically: The network status data of the device carrying the atomizing liquid is acquired at each time point. This network status data includes communication strength, network traffic, network bandwidth, signal interference, and spectrum utilization. The network performance characteristic set at the corresponding time point is analyzed, including network connectivity performance characteristic values ​​and network load interference characteristic values, specifically: Based on the communication strength, network bandwidth and spectrum utilization values ​​of the device equipped with the atomizing liquid at each time point, the network connection performance characteristics at the corresponding time points are analyzed. Based on the network traffic and signal interference values ​​of the device carrying the atomizing liquid at each time point, the network load interference characteristic values ​​at the corresponding time points are analyzed. Based on the network performance feature set of the device equipped with the atomizing liquid at each time point, the network adaptation feature value at the corresponding time point is analyzed and uploaded and stored. The cloud-based spectral processing module is used to input the spectral time series data of the set atomized liquid stored in the cloud platform into the pre-trained atomization spectral analysis model to analyze the spectral spraying mapping time series feature set of the set atomized liquid, including the light absorption spraying mapping time series feature set and the scattering spraying mapping time series feature set. The absorption spectrum time series data includes the wavelength value of each absorption data point at each time point, as well as the corresponding absorptivity and transmittance values. The scattering spectrum time series data includes the wavelength value of each scattering data point at each scattering angle at each time point, as well as the corresponding scattering intensity value. The light absorption spraying mapping time series feature set includes the absorption-transmission balance feature value, light absorption response intensity feature value, and light absorption symmetry feature value at each time point. The scattering spraying mapping time series feature set includes the scattering concentration feature value, particle size distribution feature value, and scattering distribution covariance feature value at each time point. The atomization spectral analysis model includes an absorption spectral analysis subnetwork and a scattering spectral analysis subnetwork. The specific steps for analyzing and setting the spray mapping spectral time-series feature set of the atomized liquid are as follows: In the absorption spectrum analysis sub-network of the atomization spectrum analysis model, the time series data of the absorption spectrum of the set atomized liquid are received, and the absorption time series features are extracted and analyzed to obtain the light absorption spraying mapping time series feature set of the set atomized liquid. In the scattering spectrum analysis subnetwork of the atomization spectrum analysis model, the time-series data of the scattering spectrum of the set atomized liquid are received, and the scattering time-series features are extracted and analyzed to obtain the scattering spraying mapping time-series feature set of the set atomized liquid. The cloud-based spectral analysis module is used to analyze the atomization absorption synergistic characteristic value and atomization dispersion quality characteristic value of the set atomized liquid based on the spectral spraying mapping time sequence characteristic set of the set atomized liquid. The cloud-based control feedback module is used to perform joint spraying control of the set atomized liquid based on the atomization absorption synergy characteristic value and the atomization dispersion quality characteristic value.

2. The cloud-based collaborative intelligent analysis system for atomized liquid spectral data according to claim 1, characterized in that, The absorption spectral analysis subnetwork includes an absorption input layer, an absorption extraction layer, and an absorption output layer. The specific steps for analyzing and setting the time-series feature set of the light absorption spraying mapping of the atomized liquid are as follows: In the absorption input layer of the absorption spectral analysis subnetwork, the time-series data of the absorption spectrum of the set atomized liquid is received and normalized. In the absorption extraction layer of the absorption spectral analysis subnetwork, the light absorption reaction time feature vector of the set atomized liquid is extracted based on the time series data of the absorption spectrum of the set atomized liquid after normalization. In the absorption output layer of the absorption spectrum analysis subnetwork, the light absorption reaction time sequence feature vector of the atomized liquid is set, and the light absorption spraying mapping time sequence feature set of the atomized liquid is output.

3. The cloud-based collaborative intelligent analysis system for atomized liquid spectral data according to claim 1, characterized in that, The scattering spectrum analysis subnetwork includes a scattering input layer, a scattering spray feature extraction layer, and a scattering output layer. The specific steps for analyzing and setting the scattering spray mapping time-series feature set of the atomized liquid are as follows: In the scattering input layer of the scattering spectrum analysis subnetwork, the time-series data of the scattering spectrum of the set atomized liquid is received and normalized. In the scattering spray feature extraction layer of the scattering spectrum analysis subnetwork, the scattering response time-series feature vector of the set atomized liquid is extracted based on the normalized scattering spectrum time-series data of the set atomized liquid. In the scattering output layer of the scattering spectrum analysis subnetwork, the scattering spraying mapping time sequence feature set of the atomized liquid is output based on the set scattering response time sequence feature vector of the atomized liquid.

4. The cloud-based collaborative intelligent analysis system for atomized liquid spectral data according to claim 1, characterized in that, The specific steps for analyzing and setting the synergistic characteristic value of atomization absorption of the atomizing fluid are as follows: Based on the absorption-transmission equilibrium characteristic value, light absorption response intensity characteristic value and light absorption symmetry characteristic value of the atomizing liquid at each time point, the initial atomization absorption synergistic characteristic value at the corresponding time point is analyzed. The initial atomization absorption synergistic characteristic value of the set atomizing liquid at each time point is processed by moving average to obtain the atomization absorption synergistic characteristic value of the set atomizing liquid.

5. The cloud-based collaborative intelligent analysis system for atomized liquid spectral data according to claim 1, characterized in that, The specific steps for analyzing and setting the atomization dispersion quality characteristic values ​​of the atomizing fluid are as follows: Based on the scattering concentration characteristic value, particle size distribution characteristic value and scattering distribution covariance characteristic value of the atomizing liquid at each time point, the initial atomization dispersion quality characteristic value at the corresponding time point is analyzed. The initial atomization dispersion quality characteristic values ​​of the set atomizing liquid at each time point are processed by moving average to obtain the atomization dispersion quality characteristic values ​​of the set atomizing liquid.

6. The cloud-based collaborative intelligent analysis system for atomized liquid spectral data according to claim 1, characterized in that, The specific steps for joint spray control of the set atomized liquid based on the synergistic characteristic value of atomization absorption and the characteristic value of atomization dispersion quality are as follows: The atomization absorption synergistic characteristic value and atomization dispersion quality characteristic value of the atomizing liquid were normalized. The normalized atomization absorption synergistic characteristic value and atomization dispersion quality characteristic value of the set atomizing liquid are compared with several preset atomization spraying adjustment ranges for judgment and analysis. Based on the judgment and analysis results, preset spraying control measures are selected.

Images