Sludge detection method and system based on multi-dimensional data fusion

Abstract

The invention provides a sludge detection method and system based on multi-dimensional data fusion. The method comprises the steps of multi-modal sensor array preparation, data preprocessing, timestamp alignment, feature extraction, multi-dimensional data fusion, adaptive calibration and the like. By integrating a plurality of sensors such as a near-infrared spectrometer and an ultrasonic detector and combining a deep learning algorithm, real-time monitoring and comprehensive analysis of multi-dimensional characteristics such as sludge moisture content and organic matter content are realized. The method can overcome the defects of limited single index capability, data isolation and analysis lagging of a traditional detection method, significantly improves the detection efficiency and reliability, reduces the cost, is suitable for large-scale industrial application, and provides a scientific basis for sewage treatment process optimization and resource recovery.

Description

Technical Field

[0001] This invention belongs to the field of environmental monitoring and intelligent analysis technology, specifically relating to a sludge detection method and system based on multidimensional data fusion. Background Technology

[0002] Sludge testing, as a key technological support for wastewater treatment process optimization and resource recovery, is widely used in municipal and industrial wastewater treatment. With the upgrading of environmental governance needs, traditional testing methods have evolved from manual sampling to online monitoring. However, the existing technological system still relies primarily on single-sensor detection, mainly using near-infrared spectrometers, pressure gauges, and other equipment to obtain isolated parameters such as moisture content and organic matter content. Specifically, this technological system covers the entire process from data acquisition to offline analysis, including key steps such as sample pretreatment, spectral scanning, and mechanical testing. Near-infrared spectroscopy identifies chemical components through molecular vibrational characteristics, while ultrasonic testing assesses pore structure based on sound wave propagation characteristics. However, these technologies typically operate independently and lack the ability to collaboratively analyze multi-dimensional data.

[0003] Existing sludge detection methods directly analyze data from a single sensor without establishing a fusion mechanism for multi-source heterogeneous data. This may lead to insufficient basis for process decisions or assessment results lagging behind actual operating conditions, thus affecting wastewater treatment efficiency and resource recovery value. Specifically, while traditional near-infrared spectrometers can quickly acquire chemical composition information, their detection results are easily affected by environmental temperature and humidity and cannot be correlated with sludge mechanical property data. Although ultrasonic detectors can analyze internal structures, they are difficult to form a closed-loop feedback with parameters such as moisture content when used independently. Based on this, existing technologies usually use offline laboratory analysis to supplement data, but this method has limitations such as long detection cycles (e.g., several hours to several days) and high costs, resulting in a lack of real-time data support for wastewater treatment plants when adjusting processes. Summary of the Invention

[0004] The present invention aims to at least partially solve one of the technical problems in the related art.

[0005] Therefore, the first objective of this invention is to propose a sludge detection method based on multidimensional data fusion.

[0006] The main objective of this invention is to provide a sludge detection method and system based on multidimensional data fusion. This method and system, by integrating multiple detection techniques and data analysis algorithms, addresses the shortcomings of existing sludge detection methods in terms of multidimensional characteristic monitoring, real-time performance, accuracy, and intelligence, thereby improving the efficiency and reliability of sludge detection. This invention provides a novel solution to the problems of limited single-index detection capabilities, isolated data acquisition, and delayed analysis in existing sludge detection technologies. This solution enables real-time monitoring and comprehensive analysis of multidimensional characteristics of sludge, such as moisture content, organic matter content, mechanical strength, and calorific value, without relying on traditional laboratory instruments. Its core feature lies in constructing an intelligent data processing framework that dynamically fuses and deeply analyzes raw data acquired from multiple sensors, thereby achieving a comprehensive characterization of sludge properties. Furthermore, the system features a modular hardware architecture, facilitating expansion and maintenance, and is suitable for the detection needs of different types of sludge.

[0007] The second objective of this invention is to propose a sludge detection system based on multidimensional data fusion.

[0008] The third objective of this invention is to propose a sludge detection device based on multidimensional data fusion.

[0009] The fourth objective of this invention is to provide a computer device.

[0010] The fifth objective of this invention is to provide a non-transitory computer-readable storage medium.

[0011] To achieve the above objectives, a first aspect of the present invention proposes a sludge detection method based on multidimensional data fusion, comprising: S1 synchronously collects raw data on the chemical composition, internal pore structure, compressive strength, and environmental parameters of sludge through a multimodal sensor array; S2, the original data is denoised and normalized, and timestamps are added to the data from each sensor. The data from different sources are unified to the same time reference through a timestamp alignment algorithm to form an integrated dataset. S3, input the integrated dataset into the deep learning model to extract key feature variables, including the percentage of water content, the proportion of organic matter, and the particle size distribution range in the sludge; S4. Based on the key feature variables, a multidimensional data fusion algorithm is applied to perform correlation analysis to generate a comprehensive analysis result that includes the multidimensional characteristics of sludge. S5 compares the current comprehensive analysis results with the historical comprehensive analysis results, triggers the adaptive calibration module, and automatically adjusts the sensor sensitivity and algorithm parameters.

[0012] In one embodiment of the present invention, S1 includes: S11, the multimodal sensor array adopts a distributed installation strategy, in which the near-infrared spectrometer is fixed directly above the sludge sample and about 30 cm away from the sample surface, and the ultrasonic detector is embedded on both sides of the sludge sample with coupling agent applied between the probe and the sample contact surface. S12, a pressure sensor is installed at the bottom of the sludge sample to measure compressive strength, and a temperature and humidity sensor is placed in the environment around the sludge sample to collect ambient temperature and humidity data.

[0013] In one embodiment of the present invention, S2 includes: S21 uses a signal conditioning module to filter the raw data, eliminating noise introduced by electromagnetic interference and mechanical vibration in the sensor's working environment; S22, normalizes the data collected by different sensors to the same numerical range using a normalization formula to eliminate dimensional differences.

[0014] In one embodiment of the present invention, S4 further includes: S41, by inputting the percentage of moisture content and the proportion of organic matter into the calorific value prediction model, the calorific value prediction model adopts the formula... Calculate the calorific value of the sludge, where Calorific value, This represents the percentage of moisture content. The proportion of organic matter. , , These are model parameters; S42, Generate a mechanical strength assessment sub-report based on the correlation between particle size distribution range and porosity, wherein the correlation is expressed by the formula... Characterization, in which For compressive strength, The average particle size, Porosity This is an empirical coefficient.

[0015] In one embodiment of the present invention, the method further includes: S6 will comprehensively evaluate the sludge calorific value in the assessment report. Compared with the preset calorific value threshold To make a comparison, if This will generate recommendations for the resource utilization of sludge, including biomass fuel ratio parameters.

[0016] To achieve the above objectives, a second aspect of the present invention provides a sludge detection system based on multidimensional data fusion, comprising: The system comprises a multimodal sensor array, a central processing unit, a fiber optic network, and an adaptive calibration module. The multimodal sensor array, consisting of a near-infrared spectrometer, an ultrasonic detector, a pressure sensor, and a temperature and humidity sensor, is used to collect raw data from sludge samples. The fiber optic network connects the multimodal sensor array to the central processing unit, enabling efficient data transmission. The central processing unit performs data preprocessing, timestamp alignment, feature extraction, multidimensional data fusion, and result output. The adaptive calibration module adjusts sensor sensitivity and algorithm parameters based on historical data comparisons.

[0017] The beneficial effects of the embodiments of the present invention are as follows: The aforementioned sludge detection method and system based on multidimensional data fusion, through the combination of a multimodal sensor array and intelligent algorithms, achieves comprehensive monitoring of the multidimensional characteristics of sludge. This method overcomes the limitation of single-index detection capabilities in existing technologies, providing richer data support; it not only solves the problems of data isolation and analysis lag in the sludge detection process, but also significantly improves detection efficiency and result reliability. By real-time monitoring of key indicators such as sludge moisture content, organic matter content, and mechanical strength, it provides a scientific basis for wastewater treatment process optimization and resource recovery. Compared with traditional laboratory instrument analysis, the method and system of this invention are lower in cost and simpler to operate, meeting the needs of large-scale industrial applications.

[0018] The method and system of this invention have good scalability and adaptability. Through modular hardware design and adaptive calibration mechanism, they can flexibly meet the detection needs of different types of sludge, providing strong technical support for the development of modern wastewater treatment technology.

[0019] To achieve the above objectives, a fourth aspect of this application provides a computer device, including a processor and a memory; wherein the processor runs a program corresponding to the executable program code stored in the memory, for implementing the sludge detection method based on multidimensional data fusion as described in the first aspect embodiment.

[0020] To achieve the above objectives, the fifth aspect of this application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the sludge detection method based on multidimensional data fusion as described in the first aspect embodiment.

[0021] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0022] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart of a sludge detection method based on multidimensional data fusion according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the overall structure of a sludge detection system based on multidimensional data fusion according to an embodiment of the present invention; Figure 3 This is an architecture diagram of a sludge detection method based on multidimensional data fusion according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the working principle of the adaptive calibration module according to an embodiment of the present invention; Figure 5 This is a schematic diagram of a multimodal sensor array arrangement according to an embodiment of the present invention; Figure 6 This is a structural diagram of a sludge detection device based on multidimensional data fusion according to an embodiment of the present invention; Figure 7 It is a computer device according to an embodiment of the present invention. Detailed Implementation

[0023] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0024] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0025] The following description, with reference to the accompanying drawings, describes a sludge detection method and system based on multidimensional data fusion according to an embodiment of the present invention.

[0026] Example 1 Figure 1 This is a flowchart of a sludge detection method based on multidimensional data fusion according to an embodiment of the present invention, such as... Figure 1 As shown, it includes: S1 synchronously collects raw data on the chemical composition, internal pore structure, compressive strength, and environmental parameters of sludge through a multimodal sensor array.

[0027] In some implementations, the multimodal sensor array consists of a near-infrared spectrometer, an ultrasonic detector, a pressure sensor, and temperature and humidity sensors, corresponding to the chemical composition, internal structure, mechanical properties, and environmental conditions of the sludge, respectively. The near-infrared spectrometer emits light signals in the near-infrared band (typically 780–2500 nm), which are absorbed by molecular vibrations in components such as water and organic matter in the sludge. The reflected signals are received by a photodetector and converted into spectral data for subsequent component analysis. Its installation position is fixed directly above the sludge sample, approximately 30 cm above the sample surface, to ensure the coverage of the spectral signal and the stability of the acquisition.

[0028] The ultrasonic detectors are embedded on both sides of the sludge sample, with a coupling agent (such as glycerin or water-based coupling agent) applied between the probe and the sample contact surface to reduce air impedance and signal attenuation during sound wave propagation. Their operating frequency is typically in the range of 20–100 kHz. By measuring the echo time difference and signal strength, structural parameters such as porosity and particle size distribution within the sludge can be derived. A pressure sensor is installed at the bottom of the sludge sample to measure its mechanical response during compression, outputting a pressure-deformation curve to assess compressive strength. Temperature and humidity sensors are placed in the environment surrounding the sludge sample to collect temperature and humidity data, compensating for the influence of environmental factors on the test results.

[0029] Optionally, each sensor is connected to the central processing unit via a fiber optic network to achieve high-speed, low-latency data transmission. During synchronous acquisition, all sensors are activated under the control of a unified clock signal to ensure that the acquired data is aligned on the time axis. The acquisition frequency can be set according to actual needs; for example, the near-infrared spectrometer performs periodic scanning at 10 Hz, the ultrasonic detector performs pulse emission at 50 Hz, the pressure sensor performs continuous sampling at 100 Hz, and the temperature and humidity sensor acquires environmental parameters at 1 Hz.

[0030] Furthermore, S1 includes: S11, the multimodal sensor array adopts a distributed installation strategy, in which the near-infrared spectrometer is fixed directly above the sludge sample and about 30 cm away from the sample surface, the ultrasonic detector is embedded on both sides of the sludge sample and the contact surface between the probe and the sample is coated with coupling agent.

[0031] In some implementations, this step employs a modular support structure, symmetrically arranging the spectrometer's transmitting and receiving modules to ensure a stable optical path and perpendicular illumination of the sample surface. An installation height of 30 cm is one of the experimentally validated optimal parameters, balancing the penetration depth and coverage area of ​​the spectral signal. At this height, near-infrared light can effectively penetrate the sludge surface while maximizing the spot coverage area, thereby improving spatial resolution and data representativeness. Specifically, this height allows the spectrometer's detection area to cover a circular region with a diameter of approximately 40–50 cm, meeting the typical size requirements of sludge samples on the detection platform.

[0032] Specifically, the sampling frequency of the near-infrared spectrometer is typically set to 10 Hz to ensure sufficient temporal resolution during sludge state changes. The spectral resolution is generally set to 8 cm⁻¹. -1 This enhances the ability to identify key components such as organic matter and moisture. Furthermore, to reduce ambient light interference, the system can optionally employ a light shield or background light compensation algorithm to ensure the purity of the collected data.

[0033] S12, a pressure sensor is installed at the bottom of the sludge sample to measure compressive strength, and a temperature and humidity sensor is placed in the environment around the sludge sample to collect ambient temperature and humidity data.

[0034] In some implementations, the pressure sensor employs a high-precision strain gauge pressure sensor, mounted on the bottom support structure of the sludge sample to ensure accurate sensing of the sample's deformation response when vertical pressure is applied. The sensor's range is typically set to... To accommodate the mechanical changes of sludge samples with different moisture contents and organic matter contents during the compression process, the sampling frequency was [number missing]. This ensures that instantaneous pressure responses can be captured during dynamic loading. The pressure sensor is connected to the central processing unit via a fiber optic network to ensure stable data transmission and low latency.

[0035] The temperature and humidity sensor uses an industrial-grade digital temperature and humidity probe, which is placed in the air environment surrounding the sludge sample, approximately [distance missing] from the sample surface. To avoid measurement bias caused by direct contact with the sample, the sensor's temperature measurement range is... The humidity measurement range is The precisions are respectively and Its sampling frequency is This ensures continuous monitoring of environmental parameters. Temperature and humidity sensors, through a distributed installation strategy, work in conjunction with pressure sensors to provide environmental background information for subsequent multi-dimensional data fusion.

[0036] Specifically, the pressure sensor's output signal is an analog voltage signal, which needs to be converted into a digital signal by an analog-to-digital converter (ADC), with a sampling accuracy of [insert accuracy here]. This ensures high-resolution data. The temperature and humidity sensor outputs standard I²C or SPI digital signals, facilitating direct connection to the central processing unit for data processing. Both sensors must meet synchronization requirements during data acquisition; that is, all sensor data should have a unified timestamp during the detection process to support subsequent timeline alignment and feature extraction.

[0037] S2, the original data is denoised and normalized, and timestamps are added to the data from each sensor. The data from different sources are unified to the same time reference through a timestamp alignment algorithm to form an integrated dataset.

[0038] In some implementations, noise reduction is primarily achieved through digital filters and wavelet transform techniques. For spectral signals acquired by a near-infrared spectrometer, a low-pass filter (with a cutoff frequency of...) is used. High-frequency noise is removed, and wavelet threshold denoising methods (such as Daubechies wavelet basis functions, with a decomposition level of 5) are used to locally suppress non-stationary noise. For the echo signal acquired by the ultrasonic detector, a combination of median filtering and adaptive noise cancellation algorithm is used to eliminate impulse noise and environmental interference caused by mechanical vibration. The signals from the pressure sensor and temperature and humidity sensor are filtered by moving average (window length is...). Smoothing is performed to improve data stability.

[0039] Normalization is performed using the Min-Max Normalization method, mapping the data collected by each sensor to a uniform numerical range [0, 1]. The specific formula is as follows:

[0040] in, The original data, and These are the minimum and maximum values ​​of the sensor data within the current detection period, respectively. Normalization ensures consistency in numerical scale across different sensor data, providing standardized input for subsequent timestamp alignment and feature extraction.

[0041] Furthermore, the timestamp alignment algorithm employs a mechanism based on timestamp interpolation and synchronization compensation. Each sensor appends a high-precision timestamp (accuracy 100%) to the data it collects. The central processing unit (CPU) sorts and interpolates the data using a time-axis alignment algorithm to eliminate time shifts caused by differences in sensor sampling frequencies or transmission delays. For example, the sampling frequency of a near-infrared spectrometer is... The sampling frequency of the ultrasonic detector is There is a relationship between the two. The frequency difference. Using a timestamp alignment algorithm, the system can interpolate and align two sets of data along the time dimension, ensuring that at any given time point... All sensor data can be synchronized and matched to form an integrated dataset under a unified time reference.

[0042] Furthermore, S2 includes: S21 uses a signal conditioning module to filter the raw data, eliminating noise introduced by electromagnetic interference and mechanical vibration in the sensor's working environment.

[0043] In some implementations, the signal conditioning module employs a multi-stage filtering structure, including a low-pass filter (LPF), a band-pass filter (BPF), and an adaptive noise canceller (ANC). The low-pass filter is used to suppress high-frequency noise, and its cutoff frequency is typically set at... Within a certain range, a bandpass filter is used to preserve the effective frequency band of the sensor signal. For example, the echo signal acquired by an ultrasonic detector is typically located within a certain frequency band. Within this range, the passband range of the BPF should therefore match this frequency band. Furthermore, to address periodic interference caused by mechanical vibration, the system can optionally introduce an adaptive noise canceller. This canceller acquires environmental vibration signals through a reference channel and uses a minimum mean square error (LMS) algorithm for real-time noise cancellation, with an update step size... Typically set to This is to ensure a balance between convergence speed and stability.

[0044] Furthermore, the performance metrics of filtering include signal-to-noise ratio (SNR) improvement, noise power rejection ratio (NSR), and signal fidelity. In this system, after filtering, the SNR of the original data can be improved to [value missing]. NSR reached This significantly improves data quality. Signal fidelity is quantified and evaluated using mean square error (MSE), requiring... This is to ensure that the deviation between the filtered signal and the original valid signal is within an acceptable range.

[0045] S22, normalizes the data collected by different sensors to the same numerical range using a normalization formula to eliminate dimensional differences.

[0046] In some implementations, the normalization process uses a linear normalization formula. ,in This represents the raw data value collected by a certain sensor. and These represent the minimum and maximum values ​​of the data collected by the sensor during the current detection period, respectively. This formula linearly maps the original data to the [0,1] interval, making the data from different sensors comparable on a numerical scale, thereby avoiding model training bias or fusion error caused by differences in units.

[0047] Optionally, in a practical system, the normalization operation is performed in the signal conditioning module of the central processing unit 2. This module first filters the raw data to remove noise caused by electromagnetic interference or mechanical vibration. Then, it calculates the minimum and maximum values ​​for each sensor channel within its current detection period and applies the aforementioned normalization formula for standardization. For example, the raw values ​​of the reflected light intensity data acquired by the near-infrared spectrometer 3 may be distributed in the range [0, 255], while the echo time difference data acquired by the ultrasonic detector 4 may be distributed in the range [10, 100] microseconds. Through normalization, these data will be unified to the [0,1] interval, facilitating subsequent feature extraction and multi-dimensional data fusion algorithms.

[0048] Furthermore, the accuracy and stability of the normalization process have a significant impact on the overall system performance. In this invention, the calculation precision of the normalization operation is typically set to floating-point format (e.g., 32-bit or 64-bit) to ensure that errors are not introduced due to numerical truncation during data fusion. In addition, to adapt to the characteristic variations of different batches of sludge samples, the normalization process can be dynamically adjusted in conjunction with the adaptive calibration module 8. For example, when the input range of a sensor shifts significantly, the system can recalculate. and This is to ensure the accuracy of the normalization results.

[0049] S3. Input the integrated dataset into the deep learning model to extract key feature variables, including the percentage of water content, the proportion of organic matter, and the particle size distribution range in the sludge.

[0050] In some implementations, the input layer of the deep learning model receives multidimensional data streams from near-infrared spectrometers, ultrasonic detectors, pressure sensors, and temperature and humidity sensors. After preprocessing and timestamp alignment, these data form a multi-channel time-series data matrix with a unified time reference, typically of dimension [missing information]. ,in Indicates the length of the time series. This indicates the number of sensor channels. The model employs a multi-layer convolutional structure, extracting spatial-temporal features from the data through local sensing and weight sharing mechanisms. For example, near-infrared spectral data has a continuous wavelength distribution in the frequency domain. The model uses a one-dimensional convolutional kernel to perform feature mapping on the spectral curve, thereby identifying the position and intensity changes of absorption peaks related to moisture and organic matter.

[0051] Furthermore, the ultrasonic detection data includes echo time difference and signal attenuation coefficient. The model extracts spatial features from this data through two-dimensional convolution to identify statistical characteristics of particle size distribution, such as mean, variance, and distribution range. Regarding parameter settings, the convolution kernel size is typically [value missing]. or The step size is 1 or 2, and the activation function uses ReLU or a variant thereof to enhance nonlinear expressiveness. The model output layer is mapped to the target feature space through a fully connected structure, outputting the percentage of water content in the sludge. Organic matter ratio and particle size distribution range Key variables, etc.

[0052] S4. Based on the key feature variables, a multidimensional data fusion algorithm is applied to perform correlation analysis to generate a comprehensive analysis result that includes the multidimensional characteristics of sludge.

[0053] In some implementations, multidimensional data fusion algorithms employ multilayer perceptrons (MLPs) or ensemble learning models (such as random forests and gradient boosting trees) to weight and fuse extracted feature variables. Specifically, the system takes feature vectors from near-infrared spectrometers, ultrasonic detectors, pressure sensors, and temperature and humidity sensors as input, processes them through feature normalization, and then inputs them into the fusion model. Internally, the model learns the correlations between various feature variables using training data; for example, there is a significant negative correlation between moisture content and organic matter ratio, while there is a positive correlation between particle size distribution and porosity. This is achieved by constructing a correlation matrix C between the features. ij The system can quantify the degree of coupling between different features, thereby improving the accuracy of the evaluation results.

[0054] Optionally, the fusion algorithm can also incorporate dimensionality reduction methods such as Principal Component Analysis (PCA) or Independent Component Analysis (ICA) to eliminate redundant features and extract the most representative composite features. For example, PCA can be used to reduce the original feature space from... Dimensions dropped to dimension( The system retains principal components with a cumulative variance contribution rate greater than 90%, thereby reducing computational complexity while ensuring information integrity. Furthermore, the system can employ a fuzzy logic reasoning mechanism to nonlinearly weight feature variables, adapting to the complex characteristics of different sludge samples.

[0055] Furthermore, the output of multidimensional data fusion includes the moisture content of the sludge. Organic matter content Calorific value (unit: Porosity Key indicators are analyzed, and heat maps, trend charts, and multi-dimensional radar charts are generated through a visualization module to intuitively display the comprehensive characteristics of the sludge. This report can serve as a scientific basis for optimizing wastewater treatment processes, adjusting dewatering equipment parameters, and formulating resource recovery strategies.

[0056] Furthermore, S4 includes: S41, by inputting the percentage of moisture content and the proportion of organic matter into the calorific value prediction model, the calorific value prediction model adopts the formula... Calculate the calorific value of the sludge, where Calorific value, This represents the percentage of moisture content. The proportion of organic matter. , , These are the model parameters.

[0057] In some implementations, this step first relies on two key feature variables output by the deep learning model in the preceding feature extraction module: and These two variables represent the mass percentage of water and the mass percentage of organic matter in the sludge, respectively, and their values ​​typically range from [value missing]. and This matches the typical composition range of sludge in actual treatment processes. Model parameters , , The value of is determined through regression analysis based on experimental data. For example, in an experiment at a wastewater treatment plant, Can be taken as , Can be taken as , Can be taken as Thus, a calorific value prediction model suitable for this type of sludge was constructed.

[0058] At the parameter level, the input variables of this model need to meet certain accuracy requirements. (Moisture content percentage) The measurement error should be controlled within Within, the proportion of organic matter The error should be less than This is to ensure the reliability of the calorific value prediction results. Model output The unit is Its prediction error should be less than 5%, which conforms to the industry standard for industrial sludge calorific value assessment (such as ASTM D5865-20).

[0059] S42, Generate a mechanical strength assessment sub-report based on the correlation between particle size distribution range and porosity, wherein the correlation is expressed by the formula... Characterization, in which For compressive strength, The average particle size, Porosity This is an empirical coefficient.

[0060] In some implementations, this step first relies on data on the internal pore structure of the sludge acquired by an ultrasonic detector, and on the particle size distribution range extracted through image recognition or particle size analysis algorithms. The system then calculates the particle size distribution using statistical analysis methods. For example, an arithmetic mean, geometric mean, or weighted average can be used, the specific choice depending on the dispersion of the particle distribution. Porosity The accuracy of porosity can be further verified by inverting the calculation using the attenuation characteristics and propagation time difference of the ultrasonic echo signal, combined with the known sludge density model.

[0061] Furthermore, the empirical coefficient The determination needs to combine historical data with laboratory calibration results. In some implementations, The compressive strength of multiple samples can be regressed using the least squares method in relation to particle size and porosity to obtain the optimal fit. For example, for dewatered sludge with a moisture content below 70%, It can be set to 1.2; however, for primary sludge with a high water content, It can be adjusted to 0.8 to reflect the differences in the impact of different sludge types on mechanical strength.

[0062] S5 compares the current comprehensive analysis results with the historical comprehensive analysis results, triggers the adaptive calibration module, and automatically adjusts the sensor sensitivity and algorithm parameters.

[0063] In some implementations, the adaptive calibration module first retrieves historical datasets with similar characteristics to the current sample from a cloud database, including but not limited to key feature variables such as moisture content, organic matter content, and particle size distribution. Subsequently, the system uses a sliding window mechanism to compare and analyze the current detection data with historical data, calculating their deviation values ​​across multiple dimensions. For example, if the current sludge sample's moisture content... Compared with historical average The difference exceeds the preset threshold That is, satisfying This will trigger the calibration and adjustment process.

[0064] Furthermore, based on the degree of deviation, the system employs a proportional-integral-derivative (PID) control algorithm or a gradient descent-based parameter optimization method to dynamically correct the sensor's sensitivity coefficient. Taking a near-infrared spectrometer as an example, its sensitivity can be adjusted by changing the gain coefficient of the spectral signal. To achieve this, the formula is: ,in As the reference gain, To adjust the coefficient, This represents the deviation between the current and historical moisture content. This adjustment mechanism ensures that the system maintains stable detection accuracy even when sample characteristics fluctuate.

[0065] Specifically, the adaptive calibration module supports multi-sensor collaborative calibration, such as the calibration parameters of temperature and humidity sensors. and It can dynamically compensate for environmental changes, and its adjustment range is usually limited to... To avoid system instability caused by overfitting, the system supports custom calibration cycles, such as automatically performing a calibration operation every 24 hours or each sample batch change, ensuring detection consistency over long-term operation.

[0066] S6 will comprehensively evaluate the sludge calorific value in the assessment report. Compared with the preset calorific value threshold To make a comparison, if This will generate recommendations for the resource utilization of sludge, including biomass fuel ratio parameters.

[0067] From a technical implementation perspective, the calorific value of sludge... The calorific value is calculated by correlating organic matter content data collected by a near-infrared spectrometer with pore structure information obtained by an ultrasonic detector. Specifically, the system employs a regression model based on calorific value prediction, inputting characteristic variables such as organic matter ratio, moisture content, and particle size distribution into the model, and outputting the theoretical calorific value of the sludge. The unit is Preset calorific value threshold This is set based on the minimum combustion efficiency requirements for biomass fuels, and the typical value range is [value range missing]. The specific value can be adjusted according to the performance of the combustion equipment in the actual application scenario.

[0068] If detected If the system determines that the sludge has the potential for resource utilization, it will further generate a biomass fuel blending recommendation. Blending parameters This parameter is used to quantify the degree of matching between the calorific value of sludge and fuel standards. It reflects the proportion of sludge that can be used directly as biomass fuel without the addition of auxiliary fuels, facilitating the optimization of mixing ratios in subsequent processes.

[0069] This invention discloses a sludge detection method based on multidimensional data fusion, which can realize real-time synchronous monitoring and intelligent fusion analysis of multidimensional characteristics such as sludge moisture content and organic matter content, improve detection accuracy and response speed, reduce detection costs, and is suitable for industrial applications of wastewater treatment process optimization and resource recycling.

[0070] Example 2 The following description, in conjunction with the accompanying drawings, details an embodiment of the sludge detection system based on multidimensional data fusion according to the present invention.

[0071] In this embodiment, the overall system structure is as follows: Figure 2 As shown, the system includes a multimodal sensor array 1, a central processing unit 2, an optical fiber network 7, and an adaptive calibration module 8. The multimodal sensor array 1 consists of a near-infrared spectrometer 3, an ultrasonic detector 4, a pressure sensor 5, and a temperature and humidity sensor 6. Each sensor is connected to the central processing unit 2 via the optical fiber network 7, forming a complete data acquisition and processing system. In practical applications, sludge samples are placed on a testing platform, and a multimodal sensor array is arranged around the sludge sample using a distributed installation strategy. A near-infrared spectrometer is fixed directly above the sludge sample, approximately 30 cm from the sample surface, to ensure maximum spectral signal coverage. The primary function of the near-infrared spectrometer is to capture information about the chemical composition of the sludge sample, such as organic matter content and moisture distribution. Ultrasonic detectors are embedded on both sides of the sludge sample, with coupling agent applied between the probe and the sample contact surface to reduce signal attenuation. The ultrasonic detectors are used to acquire information about the internal pore structure of the sludge, providing fundamental data for subsequent analysis of the sludge's mechanical strength. A pressure sensor is installed at the bottom of the sludge sample to measure its compressive strength, thereby assessing its mechanical properties. Temperature and humidity sensors are placed in the environment surrounding the sludge sample to monitor changes in environmental conditions such as temperature and humidity; this data is of significant reference value for analyzing sludge characteristics. All sensors are connected to the central processing unit (CPU) via a fiber optic network. The fiber optic network design avoids signal interference issues that can arise from wireless transmission, while ensuring high efficiency and stability in data transmission. After receiving the raw data from the sensors, the CPU first performs preprocessing. This preprocessing includes noise removal and normalization. Noise interference mainly originates from electromagnetic interference and mechanical vibration in the sensor's operating environment. Filtering the raw data using a signal conditioning module effectively improves the accuracy of subsequent analysis. Normalization unifies the data collected by different sensors to the same numerical range, facilitating subsequent timestamp alignment and feature extraction operations. Timestamp alignment is a crucial step in data preprocessing, such as... Figure 3As shown, this step uses a timeline alignment algorithm to unify data from different sources under the same time reference, forming a preliminary integrated dataset. The core of the timestamp alignment algorithm is to add a timestamp to each group of sensor data and adjust the data's order according to the timestamp, ensuring that data from different sensors remain consistent in the time dimension. For example, there may be a slight time difference between chemical composition data collected by a near-infrared spectrometer and pore structure data collected by an ultrasonic detector. The timestamp alignment algorithm can eliminate this difference, thus laying the foundation for subsequent multidimensional data fusion. After timestamp alignment, the data is fed into a deep learning model for feature extraction. The deep learning model employs a convolutional neural network architecture, enabling it to extract key feature variables from complex datasets. For example, analysis of near-infrared spectral data can extract the percentage of moisture and the proportion of organic matter in the sludge; analysis of ultrasonic detection data can determine the size distribution range and porosity of sludge particles. These feature variables form the basis for subsequent multidimensional data fusion. Multidimensional data fusion is one of the core components of this invention. Its purpose is to generate a comprehensive assessment report containing the multidimensional characteristics of sludge by performing correlation analysis on the extracted feature variables. For example... Figure 3 As shown, the multidimensional data fusion algorithm employs an AI-based optimization model to deeply fuse data collected from different sensors. For example, combining chemical composition information from a near-infrared spectroscopy instrument with pore structure information from an ultrasonic detector can more accurately assess the calorific value and mechanical strength of the sludge. Furthermore, data from pressure sensors and temperature and humidity sensors are also incorporated into the fusion analysis to comprehensively characterize the multidimensional properties of the sludge. The final comprehensive evaluation report not only includes various sludge parameters but also provides visual charts to facilitate intuitive understanding of the test results. The comprehensive assessment report is displayed through a user interface and stored in a cloud database for reference in subsequent process optimization and resource recovery strategy development. In practical applications, users can access the test results via touchscreen or computer terminal. For example, in wastewater treatment plants, operators can adjust sludge dewatering process parameters based on the comprehensive assessment report, thereby improving sludge treatment efficiency. Furthermore, the introduction of the cloud database allows for the long-term storage of historical data, supporting subsequent data analysis and model optimization. To address the characteristic variations of different batches of sludge samples, this invention designs an adaptive calibration module, the working principle of which is as follows: Figure 4As shown, the adaptive calibration module automatically adjusts sensor sensitivity and algorithm parameters by comparing current detection results with historical data. For example, when the moisture content of a batch of sludge samples is detected to be significantly higher than the historical average, the adaptive calibration module triggers an adjustment mechanism to improve the sensitivity of the near-infrared spectrometer, ensuring the consistency and accuracy of the detection results. Furthermore, the adaptive calibration module can dynamically adjust the calibration parameters of the temperature and humidity sensors according to changes in environmental conditions, thereby reducing the impact of environmental factors on the detection results. In practical applications, the system of this invention exhibits excellent scalability and adaptability. For example, when processing industrial sludge, the system's detection capabilities can be expanded by adding additional sensor modules. The modular hardware design makes system maintenance and upgrades more convenient, while also reducing operating costs. Furthermore, the application of an adaptive calibration mechanism enables the system to flexibly meet the detection needs of different types of sludge, achieving efficient and accurate detection for both municipal and industrial sludge. The method and system of this invention have achieved significant results in practical applications at a wastewater treatment plant. For example, in a sludge dewatering process optimization experiment, by using this system to monitor the sludge's moisture content and organic matter content in real time, operators promptly adjusted the operating parameters of the dewatering equipment, reducing the sludge moisture content from 85% to 75%, significantly improving sludge treatment efficiency. Simultaneously, because the system can comprehensively characterize the multidimensional properties of sludge, it provides a scientific basis for subsequent resource recovery. For instance, by analyzing the calorific value data of the sludge, it can be used as a raw material for biomass fuel. Throughout the entire testing process, the collaborative relationship between the various components is crucial. The multimodal sensor array is responsible for acquiring raw data, the fiber optic network ensures efficient and stable data transmission, the central processing unit performs data preprocessing, feature extraction, and multidimensional data fusion, and the adaptive calibration module ensures the consistency and accuracy of the test results by dynamically adjusting parameter configurations. This close cooperation between components enables the system to achieve comprehensive monitoring and analysis of the multidimensional characteristics of sludge without relying on traditional laboratory instruments. To better enable those skilled in the art to fully understand and implement this invention, the specific implementation principle of this invention is further explained below with reference to a specific application scenario. First, in the initial stage of sludge sample testing, the operator places the sludge sample to be tested in the center of the testing platform. The multimodal sensor array is arranged according to... Figure 5The arrangement shown is a distributed installation around the sludge sample. A near-infrared spectrometer is fixed above the sludge sample, approximately 30 cm above the sample surface, ensuring its spectral signal covers the entire sludge surface. Near-infrared light emitted by the spectrometer interacts with the organic matter and moisture in the sludge sample; the reflected light is captured by a receiver and converted into an electrical signal, thus obtaining information on the distribution of chemical components in the sludge. Simultaneously, ultrasonic detectors are embedded on both sides of the sludge sample, with coupling agent applied between the probe and the sample contact surface to reduce signal attenuation during ultrasonic wave propagation. High-frequency sound waves emitted by the ultrasonic detectors penetrate the sludge interior; by analyzing the time difference and intensity changes of the echo signals, the pore structure characteristics of the sludge can be accurately measured. A pressure sensor is installed at the bottom of the sludge sample to monitor its compressive strength in real time, thereby assessing its mechanical properties. Temperature and humidity sensors are placed in the environment surrounding the sludge sample to collect ambient temperature and humidity data, which are crucial for subsequent analysis of sludge characteristics. Subsequently, all sensors transmit the raw data they have collected to the central processing unit via a fiber optic network. For example... Figure 2 As shown, the fiber optic network design avoids signal interference problems that may arise from wireless transmission, while ensuring high efficiency and stability of data transmission. After receiving the raw data, the central processing unit first preprocesses it. The preprocessing includes noise removal and normalization. Noise interference mainly originates from electromagnetic interference and mechanical vibration in the sensor's operating environment. Filtering the raw data through the signal conditioning module effectively improves the accuracy of subsequent analysis. Normalization unifies the data collected by different sensors to the same numerical range to facilitate subsequent timestamp alignment and feature extraction operations. Timestamp alignment is a crucial step in data preprocessing, such as... Figure 3 As shown, this step uses a timeline alignment algorithm to unify data from different sources under the same time reference, forming a preliminary integrated dataset. The core of the timestamp alignment algorithm is to add a timestamp to each group of sensor data and adjust the data's order according to the timestamp, ensuring that data from different sensors remain consistent in the time dimension. For example, there may be a slight time difference between chemical composition data collected by a near-infrared spectrometer and pore structure data collected by an ultrasonic detector. The timestamp alignment algorithm can eliminate this difference, thus laying the foundation for subsequent multidimensional data fusion. After timestamp alignment, the data is fed into a deep learning model for feature extraction. The deep learning model employs a convolutional neural network architecture, enabling it to extract key feature variables from complex datasets. For example, analysis of near-infrared spectral data can extract the percentage of moisture and the proportion of organic matter in the sludge; analysis of ultrasonic detection data can determine the size distribution range and porosity of sludge particles. These feature variables form the basis for subsequent multidimensional data fusion. Multidimensional data fusion is one of the core components of this invention. Its purpose is to generate a comprehensive assessment report containing the multidimensional characteristics of sludge by performing correlation analysis on the extracted feature variables. For example... Figure 3 As shown, the multidimensional data fusion algorithm employs an AI-based optimization model to deeply fuse data collected from different sensors. For example, combining chemical composition information from a near-infrared spectroscopy instrument with pore structure information from an ultrasonic detector can more accurately assess the calorific value and mechanical strength of the sludge. Furthermore, data from pressure sensors and temperature and humidity sensors are also incorporated into the fusion analysis to comprehensively characterize the multidimensional properties of the sludge. The final comprehensive evaluation report not only includes various sludge parameters but also provides visual charts to facilitate intuitive understanding of the test results. In practical applications, the comprehensive evaluation report is displayed through a user interface and stored in a cloud database for reference in subsequent process optimization and resource recovery strategy formulation. For example, in an actual test at a wastewater treatment plant, operators accessed the test results via a touchscreen and found that the moisture content of the sludge sample was significantly higher than expected. Based on the data support provided by the comprehensive evaluation report, operators promptly adjusted the operating parameters of the sludge dewatering equipment, reducing the sludge moisture content from 85% to 75%, significantly improving sludge treatment efficiency. Simultaneously, the system provides a scientific basis for its application as a biomass fuel feedstock through analysis of sludge calorific value data. To address the characteristic variations of different batches of sludge samples, the adaptive calibration module plays a crucial role. Figure 4 As shown, the adaptive calibration module automatically adjusts sensor sensitivity and algorithm parameters by comparing current detection results with historical data. For example, when the moisture content of a batch of sludge samples is detected to be significantly higher than the historical average, the adaptive calibration module triggers an adjustment mechanism to improve the sensitivity of the near-infrared spectrometer, ensuring the consistency and accuracy of the detection results. Furthermore, the adaptive calibration module can dynamically adjust the calibration parameters of the temperature and humidity sensors according to changes in environmental conditions, thereby reducing the impact of environmental factors on the detection results. In the above process, the collaborative relationship between the various components is crucial. The multimodal sensor array is responsible for acquiring raw data, the fiber optic network ensures the efficiency and stability of data transmission, the central processing unit completes data preprocessing, feature extraction, and multidimensional data fusion, and the adaptive calibration module ensures the consistency and accuracy of the detection results by dynamically adjusting parameter configurations. The close cooperation between the components enables the system to achieve comprehensive monitoring and analysis of the multidimensional characteristics of sludge without relying on traditional laboratory instruments. Through the above steps, the method and system of the present invention have achieved significant results in practical applications, not only improving the efficiency and reliability of sludge detection, but also providing scientific support for wastewater treatment process optimization and resource recovery.

[0072] Example 3 To achieve the above embodiments, such as Figure 6 As shown, this embodiment also provides a sludge detection device 10 based on multidimensional data fusion. The device 10 includes a multimodal sensor array module 100, a data preprocessing and time alignment module 200, a deep learning feature extraction module 300, a multidimensional data fusion analysis module 400, and a dynamic parameter adjustment module 500.

[0073] The multimodal sensor array module 100 synchronously collects raw data on the chemical composition, internal pore structure, compressive strength, and environmental parameters of the sludge through the multimodal sensor array. The data preprocessing and time alignment module 200 performs noise reduction and normalization on the raw data, adds timestamps to the data from each sensor, and unifies data from different sources to the same time reference through a timestamp alignment algorithm to form an integrated dataset. The deep learning feature extraction module 300 inputs the integrated dataset into the deep learning model to extract key feature variables, including the percentage of water content, the proportion of organic matter, and the particle size distribution range in the sludge. The multidimensional data fusion analysis module 400 applies a multidimensional data fusion algorithm to perform correlation analysis based on the key feature variables, generates a comprehensive evaluation report containing the multidimensional characteristics of sludge, and stores it in a cloud database. The dynamic parameter adjustment module 500 triggers the adaptive calibration module to automatically adjust the sensor sensitivity and algorithm parameters based on the comparison between the current detection results and historical data.

[0074] Furthermore, the aforementioned multimodal sensor array module 100 is also used for: The multimodal sensor array adopts a distributed installation strategy, in which the near-infrared spectrometer is fixed directly above the sludge sample and about 30 cm away from the sample surface, and the ultrasonic detector is embedded on both sides of the sludge sample with coupling agent applied between the probe and the sample contact surface. A pressure sensor is installed at the bottom of the sludge sample to measure its compressive strength, while temperature and humidity sensors are placed in the environment surrounding the sludge sample to collect ambient temperature and humidity data.

[0075] Furthermore, the aforementioned data preprocessing and time alignment module 200 is also used for: The signal conditioning module is used to filter the raw data to eliminate noise introduced by electromagnetic interference and mechanical vibration in the sensor's working environment; S22, normalizes the data collected by different sensors to the same numerical range using a normalization formula to eliminate dimensional differences.

[0076] Furthermore, the aforementioned multidimensional data fusion analysis module 400 is also used for: S41, by inputting the percentage of moisture content and the proportion of organic matter into the calorific value prediction model, the calorific value prediction model adopts the formula... Calculate the calorific value of the sludge, where Calorific value, This represents the percentage of moisture content. The proportion of organic matter. , , These are model parameters; S42, Generate a mechanical strength assessment sub-report based on the correlation between particle size distribution range and porosity, wherein the correlation is expressed by the formula... Characterization, in which For compressive strength, The average particle size, Porosity This is an empirical coefficient.

[0077] Furthermore, device 10 is also used for: The sludge calorific value in the comprehensive assessment report Compared with the preset calorific value threshold To make a comparison, if This will generate recommendations for the resource utilization of sludge, including biomass fuel ratio parameters.

[0078] This invention discloses a sludge detection device based on multidimensional data fusion, which can realize real-time synchronous monitoring and intelligent fusion analysis of multidimensional characteristics such as sludge moisture content and organic matter content, improve detection accuracy and response speed, reduce detection costs, and is suitable for industrial applications of wastewater treatment process optimization and resource recycling.

[0079] To implement the methods of the above embodiments, the present invention also provides a computer device, such as... Figure 7As shown, the computer device 600 includes a memory 601 and a processor 602; wherein, the processor 602 reads executable program code stored in the memory 601 to run a program corresponding to the executable program code, so as to implement the various steps of the method described above.

[0080] To implement the above embodiments, this application also proposes a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method described in the foregoing embodiments.

[0081] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0082] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A sludge detection method based on multidimensional data fusion, characterized in that, include: S1 synchronously collects raw data on the chemical composition, internal pore structure, compressive strength, and environmental parameters of sludge through a multimodal sensor array; S2, the original data is denoised and normalized, and timestamps are added to the data from each sensor. The data from different sources are unified to the same time reference through a timestamp alignment algorithm to form an integrated dataset. S3, input the integrated dataset into the deep learning model to extract key feature variables, including the percentage of water content, the proportion of organic matter, and the particle size distribution range in the sludge; S4. Based on the key feature variables, a multidimensional data fusion algorithm is applied to perform correlation analysis to generate a comprehensive analysis result that includes the multidimensional characteristics of sludge. S5 compares the current comprehensive analysis results with the historical comprehensive analysis results, triggers the adaptive calibration module, and automatically adjusts the sensor sensitivity and algorithm parameters.

2. The method as described in claim 1, characterized in that, S1 includes: S11, the multimodal sensor array adopts a distributed installation strategy, in which the near-infrared spectrometer is fixed directly above the sludge sample and about 30 cm away from the sample surface, and the ultrasonic detector is embedded on both sides of the sludge sample with coupling agent applied between the probe and the sample contact surface. S12, a pressure sensor is installed at the bottom of the sludge sample to measure compressive strength, and a temperature and humidity sensor is placed in the environment around the sludge sample to collect ambient temperature and humidity data.

3. The method as described in claim 1, characterized in that, The S2 includes: S21 uses a signal conditioning module to filter the raw data, eliminating noise introduced by electromagnetic interference and mechanical vibration in the sensor's working environment; S22, normalizes the data collected by different sensors to the same numerical range using a normalization formula to eliminate dimensional differences.

4. The method as described in claim 1, characterized in that, The S4 further includes: S41, by inputting the percentage of moisture content and the proportion of organic matter into the calorific value prediction model, the calorific value prediction model adopts the formula... Calculate the calorific value of the sludge, where Calorific value, This represents the percentage of moisture content. The proportion of organic matter. , , These are model parameters; S42, Generate a mechanical strength assessment sub-report based on the correlation between particle size distribution range and porosity, wherein the correlation is expressed by the formula... Characterization, in which For compressive strength, The average particle size, Porosity This is an empirical coefficient.

5. The method as described in claim 1, characterized in that, The method further includes: S6 will comprehensively evaluate the sludge calorific value in the assessment report. Compared with the preset calorific value threshold To make a comparison, if This will generate recommendations for the resource utilization of sludge, including biomass fuel ratio parameters.

6. A sludge detection system based on multidimensional data fusion, characterized in that, include: The system comprises a multimodal sensor array, a central processing unit, a fiber optic network, and an adaptive calibration module. The multimodal sensor array, consisting of a near-infrared spectrometer, an ultrasonic detector, a pressure sensor, and a temperature and humidity sensor, is used to collect raw data from sludge samples. The fiber optic network connects the multimodal sensor array to the central processing unit, enabling efficient data transmission. The central processing unit performs data preprocessing, timestamp alignment, feature extraction, multidimensional data fusion, and result output. The adaptive calibration module adjusts sensor sensitivity and algorithm parameters based on historical data comparisons.

7. The apparatus as claimed in claim 6, characterized in that, The multimodal sensor array adopts a modular hardware design, which supports the expansion of additional sensor modules to meet the detection needs of different types of sludge.

8. A sludge detection device based on multidimensional data fusion, characterized in that, include: The multimodal sensor array module synchronously collects raw data on the chemical composition, internal pore structure, compressive strength, and environmental parameters of the sludge through a multimodal sensor array. The data preprocessing and time alignment module performs noise reduction and normalization on the raw data, adds timestamps to the data from each sensor, and unifies data from different sources to the same time reference through a timestamp alignment algorithm to form an integrated dataset. The deep learning feature extraction module inputs the integrated dataset into the deep learning model and extracts key feature variables, including the percentage of water content, the proportion of organic matter, and the particle size distribution range in the sludge. The multidimensional data fusion analysis module applies a multidimensional data fusion algorithm to perform correlation analysis based on the key feature variables, generating a comprehensive analysis result that includes the multidimensional characteristics of sludge. The dynamic parameter adjustment module compares the current comprehensive analysis results with the historical comprehensive analysis results, triggers the adaptive calibration module, and automatically adjusts the sensor sensitivity and algorithm parameters.

9. A computer device, characterized in that, Including processor and memory; The processor reads executable program code stored in the memory to run a program corresponding to the executable program code, so as to implement a sludge detection method based on multidimensional data fusion as described in any one of claims 1-5.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements a sludge detection method based on multidimensional data fusion as described in any one of claims 1-5.

Images