Intelligent early warning method and system for garbage chambers

By combining the ARIMA model and the LOF algorithm, the accuracy problem of the early warning system for garbage collection rooms was solved, and efficient prediction of environmental parameters and anomaly detection of garbage collection rooms were achieved, ensuring the safety of garbage disposal and environmental hygiene.

CN119028109AInactive Publication Date: 2024-11-26SHANDONG HAIWO JIAMEI ENVIRONMENTAL ENG CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202411455242.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-18
Publication Date
2024-11-26
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing intelligent early warning system for garbage collection rooms suffers from detection errors due to external interference, resulting in a decrease in the accuracy of early warnings.

Method used

The ARIMA model is used to predict environmental parameters of garbage collection rooms, and the Local Anomaly Factor (LOF) algorithm is used for anomaly detection. By calculating and weighting the LOF scores, the sensitivity of the early warning system is adjusted to reduce false alarms and missed alarms.

Benefits of technology

This improved the accuracy of the garbage collection room early warning system, ensuring the safety and environmental hygiene of garbage disposal and reducing false alarms and missed alarms.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119028109A_ABST
    Figure CN119028109A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of data processing, in particular to an intelligent early warning method and system for garbage boxes, and the method comprises the steps: obtaining environment parameters, corresponding to each sampling point in a normalized historical period, of the garbage boxes; inputting the normalized environmental parameters of the garbage wing-rooms in the historical period into the trained model, and outputting an environmental parameter predicted value corresponding to a preset prediction period; and carrying out anomaly detection on the environmental parameter predicted values, calculating the LOF score weight of each environmental parameter predicted value of each sampling point in the prediction period, carrying out weighted average to obtain weighted LOF scores of all environmental parameter predicted values of the corresponding sampling points, and triggering an early warning mechanism when the weighted LOF scores of the sampling points are greater than a preset anomaly threshold value. According to the method, the environment parameters of the garbage chambers are predicted, and potential problems are found in time through an anomaly detection mechanism, so that corresponding prevention or response measures are taken, and safe operation of the garbage chambers is ensured.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data processing technology. More specifically, the present invention relates to an intelligent early warning method and system for a garbage compartment. Background Art

[0002] As an indispensable infrastructure in modern urban life, garbage rooms are responsible for the classification and collection of domestic waste, which is of great significance for improving the quality of life of residents, maintaining urban environmental hygiene and promoting resource recycling. However, traditional garbage rooms face a series of challenges in actual operation, especially the problems of garbage overflow and odor diffusion, which not only affect the quality of life of surrounding residents, but also may pose a threat to public health. Therefore, the introduction of a smart early warning system has become an effective way to solve these problems.

[0003] At present, the prior art, such as the patent application document with publication number CN112250481A, discloses an intelligent monitoring and safety warning system and method for anaerobic digestion of food waste. The intelligent monitoring and safety warning system for anaerobic digestion of food waste includes: an anaerobic digestion reactor for food waste, which is used to digest and treat food waste; an intelligent monitoring subsystem, which is used to detect the parameters of the anaerobic digestion reactor during the operation of the food waste, and calculate the stabilization index according to the parameters; and a safety warning subsystem, which is used to set an alarm threshold and output an alarm signal when the stabilization index reaches the alarm threshold. The patent application document obtains the stabilization index through comprehensive calculation of the digestion parameters to intelligently determine the operating stability of the digestion system, and issues a warning instruction when the stabilization index reaches the warning threshold, thereby ensuring the stable operation of the anaerobic digestion of food waste, improving the accuracy and timeliness of the warning of the anaerobic digestion reaction system, and thus improving the operating stability of the system.

[0004] However, due to the interference of external factors, there are certain degree of failure or error in the detection, which leads to the wrong calculation of stabilization index, thus reducing the accuracy of early warning. Summary of the invention

[0005] In order to solve the above-mentioned technical problem of decreased warning accuracy, the present invention provides solutions in the following aspects.

[0006] In a first aspect, a smart early warning method for a garbage compartment includes: Obtain the environmental parameters of the garbage compartment corresponding to each sampling point in the normalized historical period, wherein the environmental parameters include temperature, gas concentration and garbage capacity; input the normalized environmental parameters of the garbage compartment in the historical period into the trained model, and output the predicted values ​​of the environmental parameters corresponding to the preset prediction period; perform anomaly detection on the predicted values ​​of the environmental parameters, wherein the anomaly detection process includes: Calculate the LOF score weight of each environmental parameter prediction value of each sampling point in the prediction period, and obtain the weighted LOF score of all environmental parameter prediction values ​​of the corresponding sampling point by weighted average. When the weighted LOF score of the sampling point is greater than the preset abnormal threshold, trigger the early warning mechanism; The LOF score weight is: ; In the formula, For the forecast period The LOF score weight of the environmental parameter prediction value corresponding to each sampling point, For the forecast period The anomaly detection index of the predicted value of the environmental parameter corresponding to the sampling point, is the false alarm rate of warning in the historical period, is the warning omission rate of the historical period.

[0007] The present invention first provides predicted values ​​of environmental parameters of the garbage compartment in the future through a prediction model, but these predicted values ​​contain a certain degree of uncertainty. Therefore, anomaly detection is performed on the basis of the prediction, and the false alarm rate and missed alarm rate of the warning are comprehensively considered in the prediction process, and the LOF score is calculated and weighted, so as to adjust the sensitivity of anomaly detection, reduce false alarms and missed alarms, and improve the accuracy of warning.

[0008] Preferably, the model is an ARIMA model.

[0009] Preferably, the process of obtaining the anomaly detection index includes: Obtain the total number of warnings and the total number of correct warnings in the historical period, and use the ratio of the total number of correct warnings to the total number of warnings as the warning accuracy rate; The ratio of the early warning accuracy rate to the product of the early warning false alarm rate and the early warning missed alarm rate is used as the abnormality detection index.

[0010] During the prediction period, the predicted value of the environmental parameters at each sampling point will calculate its LOF score. By comprehensively considering multiple factors such as the warning accuracy, false alarm rate and missed alarm rate, an anomaly detection index is calculated as an important basis for setting the warning threshold to ensure that the warning system can accurately identify abnormal situations and reduce false alarms and missed alarms in actual applications.

[0011] Preferably, the process of acquiring the anomaly detection index further includes calibrating the anomaly detection index, and the calibration steps are: The prediction accuracy of the environmental parameter prediction value corresponding to each sampling point in the prediction period is calculated, and the prediction accuracy is multiplied by the anomaly detection index and normalized to obtain the anomaly detection index after the environmental parameter prediction value corresponding to the corresponding sampling point is calibrated.

[0012] The calibrated anomaly detection index can more accurately reflect the true degree of anomaly of the predicted values ​​of environmental parameters at each sampling point, thereby improving the accuracy of anomaly detection.

[0013] Preferably, the calculation process of the prediction accuracy includes: Collecting the actual value of the environmental parameter of the most recent period within the historical period, and calculating the absolute difference between the actual value and the predicted value of the environmental parameter corresponding to any sampling point in the most recent period; Calculate the mean absolute difference between the actual and predicted values ​​of the environmental parameters corresponding to all sampling points in the entire historical period; The ratio of the mean absolute difference to the absolute difference is taken as the prediction accuracy.

[0014] Based on the calculation results of prediction accuracy, we can find the performance differences of the model at different sampling points or different time periods. This helps to identify the shortcomings of the model and adjust the model parameters in a targeted manner to improve the prediction accuracy.

[0015] Preferably, the process of obtaining the anomaly detection index includes: Obtain the total number of warnings and actual anomalies in the historical period. The actual anomaly total includes the total number of correct warnings, the total number of false alarms, and the total number of missed alarms. The ratio of the total number of correct warnings to the total number of warnings is taken as the precision rate, and the ratio of the total number of correct warnings to the total number of actual anomalies is taken as the recall rate; The product of precision and recall and the sum of precision and recall are used as anomaly detection indicators.

[0016] Preferably, the calculation process of the prediction accuracy includes: During the historical period, for each sampling point, calculating the error percentage between the predicted value and the actual value of the corresponding environmental parameter; The average value of the error percentages of all sampling points is calculated, and the average value of the error percentages is taken as the prediction accuracy.

[0017] By calculating the error percentage, we can more intuitively express the degree to which the predicted value deviates from the actual value.

[0018] In a second aspect, an intelligent early warning system for a garbage compartment includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned intelligent early warning method for the garbage compartment is implemented.

[0019] The beneficial effects of the present invention are: The present invention first provides predicted values ​​of the environmental parameters of the garbage compartment in the future through a prediction model, but these predicted values ​​contain a certain degree of uncertainty. Therefore, an anomaly detection mechanism is added on the basis of the prediction, and the false alarm rate and missed alarm rate of the warning are comprehensively considered in the prediction process. The LOF score is calculated and weighted, so as to adjust the sensitivity of anomaly detection, reduce false alarms and missed alarms, and improve the accuracy of warning.

[0020] The prediction model and anomaly detection mechanism of the present invention together constitute an efficient and reliable garbage compartment early warning system, which provides a strong guarantee for the safe operation of the garbage compartment. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 It is a method flow chart of steps S1 to S3 in an intelligent early warning method for a garbage compartment in an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0023] The present invention collects historical data related to garbage compartments, and then uses a pre-trained model to predict the predicted value of data related to garbage compartments in a certain period of time in the future. On this basis, the local outlier factor (LOF) algorithm is used for anomaly detection, which can effectively identify those outliers that are significantly different from most data, and timely discover abnormal changes in environmental parameters of garbage compartments, so as to take corresponding measures, such as adjusting the frequency of garbage collection or optimizing the garbage disposal process. This method helps to improve the efficiency and effect of garbage disposal, and ensure the normal operation of garbage compartments and the sanitation and safety of the environment.

[0024] The embodiment of the present invention discloses a smart early warning method for a garbage compartment, referring to Figure 1 , including steps S1 to S3, which are specifically as follows: S1: Obtain the normalized environmental parameters of the garbage compartment corresponding to each sampling point in the historical period, wherein the environmental parameters include temperature, gas concentration and garbage capacity.

[0025] In the embodiment of the present invention, by installing sensors and other equipment, key environmental parameters such as garbage capacity, temperature, humidity and harmful gases in the garbage compartment are monitored in real time.

[0026] Specifically, there are a total of 4 garbage bins in the preset garbage room. A temperature sensor and a humidity sensor are arranged directly above each garbage bin to monitor the temperature and humidity changes in the garbage bin; a pressure sensor is arranged at the bottom of each garbage bin to monitor the weight or accumulation degree of garbage in the garbage bin; a harmful gas sensor is arranged on the top of the garbage room to monitor the concentration of harmful gases such as methane in the air.

[0027] Among them, within the historical period (in the embodiment of the present invention, various environmental parameters of the garbage compartment in the past 30 days are collected), all sensors collect data at a sampling frequency of 2 Hz, with a total of 5,184,000 sampling points, and 1 sampling point corresponds to the environmental parameters of temperature, gas concentration, humidity and garbage capacity.

[0028] In addition, after collection, these sampling points are sorted in the order of the time of collection.

[0029] Since the collected environmental parameters may be affected by factors such as electromagnetic interference during the collection process, they often contain noise. In order to improve the accuracy and reliability of the environmental parameters, it is necessary to perform denoising on the collected environmental parameters. Generally, Gaussian filtering is used to perform denoising on the collected environmental parameters.

[0030] After denoising, the environmental parameters of the garbage compartment are linearly normalized and stored in the host computer storage module through the wireless network transmission module.

[0031] S2: Input the normalized environmental parameters of the garbage compartment in the historical period into the trained model, and output the predicted values ​​of the environmental parameters corresponding to the preset prediction period.

[0032] In the embodiment of the present invention, firstly, the environmental parameters of the garbage compartment collected in the past 30 days in step S1 are used as the initial data set, and the ARIMA model is used to predict the future environmental parameters. It should be noted that this operation is to predict any one of the environmental parameters, and set the prediction period to the next week or the next day.

[0033] For example, the temperature of the garbage compartment in the past 30 days is first used as the initial data set; next, the Augmented Dickey-Fuller (ADF) test or other stationarity test methods are used to determine whether these temperature parameters are stationary. If the temperature parameters are non-stationary, differential processing is required until these temperature parameters become stable; after the temperature parameters are stable, the autocorrelation coefficient and partial autocorrelation coefficient of the temperature parameters are calculated to obtain the autocorrelation graph and partial autocorrelation graph, and the values ​​of the autoregressive terms (AR terms) and the moving average terms (MA terms) are determined based on the significant peaks in the autocorrelation graph and the partial autocorrelation graph; finally, the determined difference order, AR terms, and MA terms are used to establish the ARIMA model.

[0034] After the ARIMA model is established, the data of the past 30 days are fitted and the predicted value of the temperature parameter for the next day is generated. Similarly, the sampling points for the next day are sorted according to the above step S1.

[0035] It should be noted that since environmental parameters are constantly changing, prediction is not a one-time activity but an ongoing process, so the model needs to be constantly updated and adjusted according to the new parameters.

[0036] After each prediction cycle, the model will reselect the environmental parameter historical data of the most recent cycle as new input data. This most recent cycle can be the most recent week or other time periods set according to actual needs. After using the new input data, the model will repeat the previous prediction steps.

[0037] It should be noted that since the prediction process is continuous, when the prediction results indicate possible abnormalities, the early warning system will immediately issue an alarm until the early warning system no longer needs prediction support, that is, stops predicting.

[0038] When the residuals of the environmental parameter historical data predictions for the most recent period are lower than the residuals of the overall predictions for the historical period, it means that the recent prediction model has a better fit for the recent parameters. This usually indicates that the model can capture recent trends and patterns more accurately, so when using these recent environmental parameter historical data to predict the parameters of the next prediction period, the prediction accuracy of the environmental parameter prediction values ​​corresponding to each sampling point in the prediction period may be higher.

[0039] Specifically, the actual values ​​of the environmental parameters of the most recent period are collected during the historical period, and the absolute difference between the actual value and the predicted value of the environmental parameters corresponding to any sampling point in the most recent period is calculated; Calculate the mean absolute difference between the actual and predicted values ​​of the environmental parameters corresponding to all sampling points in the entire historical period; The ratio of the average absolute difference to the absolute difference calculated above is taken as the prediction accuracy mentioned above.

[0040] Furthermore, the relative positions between the sampling points in the most recent period and the sampling points in the prediction period are analyzed, and a weight factor is calculated to adjust the contribution of the environmental parameters corresponding to the sampling points in the most recent period to the prediction accuracy of the predicted values ​​of the environmental parameters in the prediction period.

[0041] Exemplarily, the prediction accuracy after introducing the weight factor satisfies the following relationship:

[0042] In the formula, After the weight factor is introduced The prediction accuracy of the environmental parameter prediction value corresponding to each sampling point is The first The prediction accuracy of the environmental parameter prediction value corresponding to each sampling point is For the forecast period The serial number of the sampling point, is the total number of sampling points in the most recent cycle, The most recent cycle The serial number of the sampling point, is an exponential function with base e.

[0043] The prediction accuracy of the environmental parameter prediction value is evaluated by comprehensively considering the prediction accuracy of the environmental parameter in the most recent period and the distance from the prediction period. The closer the distance between the sampling point and the sampling point with high prediction accuracy in the most recent period, the The more sampling points there are, the greater the impact they have on these high-accuracy sampling points.

[0044] In another embodiment, a simple and intuitive way is provided to evaluate the prediction accuracy of environmental parameters. The general steps are as follows: During the historical period, for each sampling point, the error percentage between the predicted value and the actual value of the environmental parameter corresponding to the sampling point is calculated, that is, ,in, is the error percentage of the sampling point, is the predicted value of the sampling point, is the actual value of the sampling point.

[0045] Then the error percentage of all sampling points in the historical period is calculated, and then the average error percentage of all sampling points in the historical period is calculated, and the average error percentage of all sampling points in the historical period is used as the prediction accuracy of the environmental parameters.

[0046] The calculated average error percentage is used as a measure of prediction accuracy. The smaller this value is, the higher the prediction accuracy of the prediction model is.

[0047] In summary, the environmental parameters of the most recent period can be used to input the trained model, and the predicted values ​​of the environmental parameters corresponding to the preset prediction period can be output with higher accuracy.

[0048] After a series of analyses above have been conducted to obtain a trained model using the environmental parameters of the most recent period as input, and outputting a more accurate result of the environmental parameter prediction values ​​corresponding to the preset prediction period, the environmental parameters of the most recent period are further used for prediction to obtain the environmental parameter prediction values ​​of the prediction period, thereby obtaining the predicted values ​​of key environmental parameters such as garbage capacity, temperature, humidity, and harmful gases corresponding to all sampling points within the prediction period.

[0049] S3: Perform anomaly detection on the predicted value of the environmental parameter to complete the early warning of the garbage compartment.

[0050] The output of the prediction model in step S2 above provides a benchmark for anomaly detection, which can reduce false positives in anomaly detection. In an embodiment of the present invention, a local anomaly factor (LOF) algorithm is used for anomaly detection. The general steps of anomaly detection are as follows: The first step is to calculate the anomaly detection index of the predicted value of the environmental parameter; the detection index satisfies the following relationship:

[0051] In the formula, It is an anomaly detection indicator for the predicted values ​​of environmental parameters within the prediction period; is the warning accuracy rate of the historical period; is the false alarm rate of warning in the historical period; is the warning omission rate of the historical period.

[0052] Among them, the total number of warnings and the total number of correct warnings in the historical period are obtained, and the ratio of the total number of correct warnings to the total number of warnings is used as the warning accuracy rate; the number of false warnings in the historical period is obtained, and the ratio of the number of false warnings to the total number of warnings is used as the warning false alarm rate; the number of real anomalies that were not detected in the historical period and the total number of actual anomalies are obtained, and the ratio of the number of real anomalies that were not detected to the total number of actual anomalies is used as the warning omission rate.

[0053] It should be noted that when or When the value of is close to 0, may become very large or unstable because the denominator ( ) will approach 0. In practical applications, it may be necessary to adjust or supplement this indicator according to specific circumstances to ensure that it can accurately reflect the performance of the prediction system.

[0054] When the warning accuracy The higher the false alarm rate, the and the missed alarm rate When low, anomaly detection indicators The higher it is, the more accurate the result of anomaly detection using the LOF (Local Outlier Factor) algorithm.

[0055] The second step is to calibrate the anomaly detection indicators; the general steps of calibration are as follows: The prediction accuracy of the environmental parameter prediction value corresponding to each sampling point in the prediction period is multiplied by the anomaly detection index and normalized to obtain the anomaly detection index after the environmental parameter prediction value corresponding to the corresponding sampling point is calibrated.

[0056] Then the anomaly detection index obtained after calibration satisfies the relationship:

[0057] In the formula, For the The anomaly detection index of the predicted value of the environmental parameter corresponding to the sampling point, For the The prediction accuracy of the environmental parameter prediction value corresponding to each sampling point is is the anomaly detection indicator of the predicted value of the environmental parameter within the prediction period, For normalization processing.

[0058] Then, the anomaly detection indicators corresponding to the predicted values ​​of key environmental parameters such as garbage capacity, temperature, humidity and harmful gases corresponding to each sampling point in the prediction period are obtained.

[0059] Through calibration, the early warning system can better handle anomalies and uncertainties, and enhance its robustness in the face of complex or unknown environments. If the prediction accuracy is high, the calibrated anomaly detection indicators will more accurately reflect the actual degree of anomaly; conversely, if the prediction accuracy is low, the calibrated anomaly detection indicators will be adjusted accordingly to avoid false positives or negatives.

[0060] In another embodiment, the process of obtaining the abnormality detection index may also be: Obtain the total number of warnings and actual anomalies in the historical period. The actual anomaly total includes the total number of correct warnings, the total number of false alarms, and the total number of missed alarms. The ratio of the total number of correct warnings to the total number of warnings is taken as the precision rate, and the ratio of the total number of correct warnings to the total number of actual anomalies is taken as the recall rate; The product of precision and recall and the sum of precision and recall are used as anomaly detection indicators.

[0061] The third step is to calculate the weighted local outlier factor (LOF) score of the environmental parameter prediction value of each sampling point within the prediction period, and trigger the early warning mechanism based on this score; the calculation steps are as follows: For each sampling point within the prediction period, the LOF score weight of each environmental parameter prediction value is calculated to satisfy the following relationship:

[0062] In the formula, For the forecast period The LOF score weight of the environmental parameter prediction value corresponding to each sampling point, For the forecast period The anomaly detection index of the predicted value of the environmental parameter corresponding to the sampling point, is the false alarm rate of the historical period, is the warning omission rate of the historical period.

[0063] By combining anomaly detection indicators and historical warning performance ( and ) to assign a reasonable LOF score weight to each environmental parameter prediction value. When it is higher (indicating that the predicted value is more abnormal), the weight will increase; when Higher When it is low (indicating that the system is prone to false positives but less prone to false negatives), the weight will also increase to reflect the system's sensitivity to anomalies.

[0064] For each sampling point, we need to take the weighted average of the LOF scores of all its environmental parameter prediction values ​​according to the weights calculated above to obtain the weighted LOF score of the sampling point. This score reflects the overall abnormality of the sampling point relative to other sampling points.

[0065] Step 4: preset an abnormal threshold. In the embodiment of the present invention, the abnormal threshold is set to 1. When the weighted LOF score of a sampling point is greater than the preset abnormal threshold, the sampling point is judged to be abnormal and an early warning mechanism is triggered. This mechanism may include sending an alarm, recording abnormal events, adjusting system parameters, etc. to deal with potential abnormal situations.

[0066] An embodiment of the present invention further discloses an intelligent early warning system for a garbage compartment, comprising a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the intelligent early warning method for a garbage compartment according to the present invention is implemented.

[0067] The system also includes other components familiar to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and will not be described in detail here.

[0068] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module or both. Any such computer storage medium may be part of a device or accessible or connectable to a device.

[0069] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0070] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. An intelligent early warning method for garbage compartments, characterized in that: include: Obtaining normalized environmental parameters of the garbage compartment corresponding to each sampling point in the historical period, the environmental parameters including temperature, gas concentration and garbage capacity; The normalized environmental parameters of the garbage compartment in the historical period are input into the trained model, and the predicted values ​​of the environmental parameters corresponding to the preset prediction period are output; Performing abnormal detection on the predicted values ​​of the environmental parameters to complete early warning of the garbage compartment; The anomaly detection process includes: Calculate the LOF score weight of each environmental parameter prediction value of each sampling point in the prediction period, and obtain the weighted LOF score of all environmental parameter prediction values ​​of the corresponding sampling point by weighted average. When the weighted LOF score of the sampling point is greater than the preset abnormal threshold, trigger the early warning mechanism; The LOF score weight is: ; In the formula, For the forecast period The LOF score weight of the environmental parameter prediction value corresponding to each sampling point, For the forecast period The anomaly detection index of the predicted value of the environmental parameter corresponding to the sampling point, is the false alarm rate of the historical period, is the warning omission rate of the historical period.

2. The intelligent early warning method for garbage compartment according to claim 1 is characterized in that: The model is an ARIMA model.

3. The intelligent early warning method for garbage compartment according to claim 1 is characterized in that: The process of obtaining the anomaly detection index includes: Obtain the total number of warnings and the total number of correct warnings in the historical period, and use the ratio of the total number of correct warnings to the total number of warnings as the warning accuracy rate; The ratio of the early warning accuracy rate to the product of the early warning false alarm rate and the early warning missed alarm rate is used as the abnormality detection index.

4. The intelligent early warning method for garbage compartment according to claim 3 is characterized in that: The process of obtaining the anomaly detection index further includes calibrating the anomaly detection index, and the calibration steps are: The prediction accuracy of the environmental parameter prediction value corresponding to each sampling point in the prediction period is calculated, and the prediction accuracy is multiplied by the anomaly detection index and normalized to obtain the anomaly detection index after the environmental parameter prediction value corresponding to the corresponding sampling point is calibrated.

5. The intelligent early warning method for garbage compartment according to claim 4 is characterized in that: The calculation process of the prediction accuracy includes: Collecting the actual value of the environmental parameter of the most recent period within the historical period, and calculating the absolute difference between the actual value and the predicted value of the environmental parameter corresponding to any sampling point in the most recent period; Calculate the mean absolute difference between the actual and predicted values ​​of the environmental parameters corresponding to all sampling points in the entire historical period; The ratio of the mean absolute difference to the absolute difference is taken as the prediction accuracy.

6. The intelligent early warning method for garbage compartment according to claim 1 is characterized in that: The process of obtaining the anomaly detection index includes: Obtain the total number of warnings and actual anomalies in the historical period. The actual anomaly total includes the total number of correct warnings, the total number of false alarms, and the total number of missed alarms. The ratio of the total number of correct warnings to the total number of warnings is taken as the precision rate, and the ratio of the total number of correct warnings to the total number of actual anomalies is taken as the recall rate; The product of precision and recall and the sum of precision and recall are used as anomaly detection indicators.

7. The intelligent early warning method for garbage compartment according to claim 4 is characterized in that: The calculation process of the prediction accuracy includes: During the historical period, for each sampling point, calculating the error percentage between the predicted value and the actual value of the corresponding environmental parameter; The average value of the error percentages of all sampling points is calculated, and the average value of the error percentages is taken as the prediction accuracy.

8. An intelligent early warning system for garbage compartments, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the intelligent early warning method for a garbage compartment according to any one of claims 1 to 7 is implemented.

Citation Information

Patent Citations

  • Intelligent monitoring and safety early warning system and method for anaerobic digestion of kitchen waste

    CN112250481A

  • Organic solid stockpiling waste monitoring method, controller, storage medium and system

    CN118034135A