Regional household garbage state monitoring method based on big data

Abstract

The invention relates to a regional household garbage state monitoring method based on big data. The regional household garbage state monitoring method comprises the following steps: step 1, acquiring household garbage data in a specified region in real time through a multi-source data acquisition system; 2, performing standardized cleaning processing on the original data, and establishing a three-dimensional data matrix containing garbage category identifiers, spatial position coordinates and timestamps; according to the method, an intelligent whole-process monitoring and management system is constructed, a solution is provided for urban garbage treatment, a multi-source data acquisition network is firstly established, intelligent garbage can sensing data, vehicle track information and community report data are integrated, real-time monitoring of the whole process of garbage generation, collection and transportation is achieved, and real-time monitoring of the whole process of garbage collection and transportation is achieved; according to the system, a standardized data processing flow and a distributed computing framework are adopted, a space-time three-dimensional data matrix is constructed, and visual presentation and trend prediction of garbage distribution are achieved through an intelligent algorithm.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of urban environment management, in particular to a regional household garbage state monitoring method based on big data. BACKGROUND

[0002] With the acceleration of urbanization and the continuous growth of population, the amount of household garbage is increasing rapidly, which brings great challenges to urban environment management and garbage disposal. The traditional garbage management mode mainly relies on manual patrol and fixed route collection and transportation, and has problems such as backward monitoring means, incomplete data collection, and untimely response, which makes it difficult to realize precise and intelligent management. In the prior art, although some cities have begun to apply Internet of Things devices such as intelligent garbage cans, there is a common phenomenon of data island, and various monitoring data cannot be effectively integrated. The garbage amount prediction mainly relies on historical experience data, and lacks dynamic consideration of external factors such as weather conditions and population flow. The data processing and analysis capability is insufficient, and it is difficult to provide real-time and accurate support for management decision-making. In addition, the existing system mostly adopts centralized processing mode in the computing architecture, and has defects such as response delay and poor scalability when facing massive monitoring data. These technical bottlenecks seriously restrict the improvement of urban garbage management efficiency, and it is urgent to build an intelligent monitoring system integrating data collection, processing, analysis and decision support to realize precise scheduling of garbage collection and transportation and optimal allocation of resources. SUMMARY

[0003] In order to solve the above problems, the application provides a regional household garbage state monitoring method based on big data.

[0004] The regional household garbage state monitoring method based on big data provided by the application adopts the following technical scheme: The regional household garbage state monitoring method based on big data comprises the following steps: Step 1, acquiring real-time household garbage data in a specified region through a multi-source data acquisition system; Step 2, standardizing and cleaning the original data, and establishing a three-dimensional data matrix containing garbage category identification, spatial position coordinates and time stamp; Step 3, using a distributed computing framework to perform clustering analysis on the preprocessed data, generating a regional garbage state heat map and a category distribution radar chart, and completing the integration processing of the data; Step 4, predicting the future garbage amount change trend by combining external influencing factors through a time series prediction model.

[0005] As the preferred technical solution of the present application, the multi-source data acquisition system includes an intelligent trash can sensor data unit, the output end of the intelligent trash can sensor data unit is electrically connected to a garbage collection truck GPS trajectory data unit, the output end of the garbage collection truck GPS trajectory data unit is electrically connected to community reporting data and a government database, and the time series prediction model includes a historical garbage volume data model, a temperature model, a precipitation model and a population flow thermal model.

[0006] As a preferred technical solution of the present application, the smart trash can sensor data unit is equipped with a pressure sensor. The smart trash can sensor data unit collects data by monitoring the changes in the trash can load in real time through the pressure sensor and establishing a weight-to-capacity correspondence curve. The smart trash can sensor data unit has a built-in infrared spectrum recognition device. The smart trash can sensor data unit collects data by using the infrared spectrum recognition device to classify the materials of the input garbage. The infrared spectrum recognition device is used to distinguish recyclables, kitchen waste and hazardous waste. The smart trash can sensor data unit is integrated with an ultrasonic ranging module. The smart trash can sensor data unit collects data by detecting the garbage filling height through the ultrasonic ranging module. When the ultrasonic ranging module detects that the garbage filling height reaches a preset threshold, a full load alarm signal is triggered. The smart trash can sensor data unit has a built-in NB-IoT communication module. The NB-IoT communication module is encapsulated and transmitted in JSON format. The transmission data of the NB-IoT communication module includes device ID, latitude and longitude coordinates and data collection time.

[0007] As the preferred technical solution of this application, in step three, more specifically, a data quality assessment system is established, missing data is interpolated and completed using the KNN algorithm, abnormal data is screened and eliminated using the 3σ principle, discrete data points are mapped to 1km×1km geographic units through spatial gridding, and the MapReduce framework of the Hadoop ecosystem is used to realize parallel processing of multi-source heterogeneous data and generate a standardized HDFS storage format.

[0008] As the preferred technical solution of this application, the time series prediction model is internally integrated with a prediction analysis module, which is used to construct an LSTM neural network model. The prediction analysis module includes an input layer, a hidden layer and an output layer. The input layer contains 72 hours of historical garbage volume time series data, temperature and humidity data obtained from the Meteorological Bureau API, and population density data parsed from mobile phone signaling. The hidden layer uses a three-layer GRU unit with a Dropout rate of 0.2. The output layer generates garbage volume forecast values ​​for the next 24 hours in different time periods. The prediction analysis module is loaded with a backpropagation algorithm, which is used to update the model weight parameters.

[0009] As a preferred technical solution of the present application, the multi-source data acquisition system further comprises a distributed data acquisition terminal cluster, an output end of the distributed data acquisition terminal cluster is electrically connected with an edge computing gateway, the edge computing gateway is built-in with data filtering rules and compression algorithms, an output end of the edge computing gateway is electrically connected with a cloud computing platform, the cloud computing platform is internally integrated with a Spark real-time analysis engine and a TensorFlow prediction model, an output end of the cloud computing platform is electrically connected with a visual interactive interface, and the visual interactive interface supports rendering a three-dimensional garbage distribution situation map and multi-dimensional data grabbing analysis through a WebGL technology.

[0010] As a preferred technical solution of the present application, the cloud computing platform adopts Kubernetes container orchestration management analysis microservices, and the cloud computing platform is internally integrated with a Redis in-memory database.

[0011] As a preferred technical solution of the present application, the visual interactive interface comprises a GIS map display layer, the GIS map display layer superimposedly displays garbage collection point distribution, real-time position of a garbage collection vehicle and a garbage landfill state, an output end of the GIS map display layer is electrically connected with a data analysis panel, an output end of the data analysis panel is electrically connected with a pre-warning management module, the data analysis panel is used to provide a month-on-month growth rate, a classification proportion pie chart and a recycling economic value column chart, the pre-warning management module automatically triggers a red alarm when detecting that a regional garbage amount exceeds a bearing capacity threshold, an output end of the pre-warning management module is electrically connected with a decision support unit, the decision support unit generates a resource allocation suggestion scheme according to a prediction result, and the resource allocation suggestion scheme comprises a vehicle scheduling route and a temporary collection point setting scheme.

[0012] In summary, the present application has the following at least based on big data regional household garbage state monitoring methods beneficial technical effects: This application provides a solution for urban waste management by constructing an intelligent full-process monitoring and management system. The method first establishes a multi-source data acquisition network, integrates smart trash can sensor data, vehicle trajectory information and community reporting data, and realizes real-time monitoring of the entire process of garbage generation, collection and transportation. The system adopts standardized data processing procedures and distributed computing frameworks to construct a three-dimensional spatiotemporal data matrix, and realizes the visualization of garbage distribution and trend prediction through intelligent algorithms. The edge-cloud collaborative computing model is adopted in the technical architecture, which not only ensures real-time response capabilities but also has powerful data analysis functions. The core prediction module is based on deep learning technology and combines multi-dimensional influencing factors to significantly improve the accuracy of garbage volume prediction. The system is equipped with intelligent early warning and decision support functions, which can automatically generate optimized scheduling plans based on real-time monitoring data and prediction results, greatly improving garbage collection and transportation efficiency. The overall solution adopts a modular design with good scalability and adaptability, and can be continuously optimized and upgraded to meet the needs of smart city development. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 This is a flow chart of the regional domestic waste status monitoring method based on big data in this application; Figure 2 This is the distribution map of the regional domestic waste status monitoring system based on big data in this application. DETAILED DESCRIPTION

[0014] The following is combined with Figure 1-2 This application is described in further detail.

[0015] See also Figure 1 ,A regional domestic waste status monitoring method based on big data, comprising the following steps: Step 1: Real-time acquisition of domestic waste data in a designated area through a multi-source data collection system; Step 2: Standardize and clean the raw data to create a three-dimensional data matrix containing garbage category identification, spatial location coordinates, and timestamps; Step three: Use a distributed computing framework to perform cluster analysis on the preprocessed data, generate regional waste status heat maps and category distribution radar maps, and complete the data integration process. A data quality assessment system is established, using the KNN algorithm to interpolate missing data and the 3σ principle to filter out abnormal data. Spatial gridding is used to map discrete data points to 1km×1km geographic units. The Hadoop ecosystem's MapReduce framework is used to parallelize multi-source heterogeneous data and generate standardized HDFS storage formats. Step 4: Use the time series prediction model combined with external influencing factors to predict the future trend of garbage volume changes; The system adopts a continuous iteration optimization strategy, the prediction model adapts to changes in data distribution through an online learning mechanism, and the data quality evaluation module dynamically corrects outliers and missing data. A / B testing and simulation verification are used to evaluate the improvement effect of the algorithm, ensuring that the updated model maintains high accuracy in actual scenarios. The operation and maintenance monitoring system tracks system performance in real time, promptly identifies and fixes potential problems. At the same time, the system supports modular expansion, allowing flexible integration of new data sources or analysis algorithms to adapt to the development needs of future smart cities.

[0016] Referring to Figure 2 , the multi-source data acquisition system includes a smart garbage can sensor data unit, the output end of the smart garbage can sensor data unit is electrically connected with a garbage collection vehicle GPS trajectory data unit, the output end of the garbage collection vehicle GPS trajectory data unit is electrically connected with community reporting data and a government database, the time series prediction model includes a historical garbage quantity data model, a temperature model, a precipitation model, and a population flow heat model.

[0017] The smart garbage can sensor data unit is internally loaded with a pressure sensor, the collection of the smart garbage can sensor data unit is realized by real-time monitoring of the load change of the garbage can through the pressure sensor and establishing a weight-to-capacity corresponding curve, the smart garbage can sensor data unit is internally provided with an infrared spectrum recognition device, the collection of the smart garbage can sensor data unit adopts the infrared spectrum recognition device to classify the material of the garbage, the infrared spectrum recognition device is used to distinguish recyclable materials, kitchen garbage, and hazardous waste, the smart garbage can sensor data unit is internally integrated with an ultrasonic ranging module, the collection of the smart garbage can sensor data unit is realized by detecting the garbage filling height through the ultrasonic ranging module, the ultrasonic ranging module detects the garbage filling height and triggers a full load alarm signal when the garbage filling height reaches a preset threshold, the smart garbage can sensor data unit is internally provided with an NB-IoT communication module, the NB-IoT communication module encapsulates and transmits in JSON format, the transmission data of the NB-IoT communication module includes a device ID, a latitude and longitude coordinate, and a data collection time; A multi-dimensional data acquisition system is formed by the smart garbage can, the GPS of the collection vehicle, and the community reporting data, the smart garbage can is equipped with a high-precision pressure sensor, an infrared spectrum analysis module, and an ultrasonic ranging device, which can monitor the weight, composition, and filling height of the garbage in real time, ensuring the comprehensiveness and accuracy of data collection. The edge computing node performs preliminary cleaning and compression on the original data and transmits it to the cloud using a lightweight protocol, significantly reducing network load. The deep integration of multi-source data not only improves the visualization of garbage status but also provides a high-quality data foundation for subsequent intelligent analysis. The system supports dynamic expansion and can flexibly access municipal databases, weather information, and other external data sources, further enhancing the real-time monitoring and reliability of predictions.

[0018] The time series prediction model integrates a prediction and analysis module, which is used to build an LSTM neural network model. The prediction and analysis module includes an input layer, a hidden layer, and an output layer. The input layer contains 72 hours of historical garbage volume time series data, temperature and humidity data obtained from the Meteorological Bureau API, and population density data analyzed from mobile phone signaling. The hidden layer uses a three-layer GRU unit with a dropout rate of 0.2. The output layer generates garbage volume forecasts for the next 24 hours by time period. The prediction and analysis module is equipped with a backpropagation algorithm, which is used to update the model weight parameters. Garbage volume prediction uses an LSTM neural network. The model input includes historical garbage volume, meteorological data, and population flow information. Time series modeling is used to capture the cyclical characteristics of garbage generation. The network structure uses multi-layer GRU units, combined with the Dropout mechanism to prevent overfitting, and introduces an attention mechanism to dynamically adjust the influence weights of external factors. The model training uses an adaptive optimization algorithm and supports online learning to continuously optimize prediction accuracy. The prediction results are not only used for short-term collection and transportation scheduling, but also can assist in the long-term planning of garbage treatment facilities. The system can predict garbage volume fluctuations during holidays or extreme weather, adjust collection and transportation resources in advance, and avoid garbage accumulation. The model's interpretability analysis helps managers understand the prediction logic and enhance the credibility of decision-making.

[0019] The multi-source data collection system also includes a distributed data collection terminal cluster. The output end of the distributed data collection terminal cluster is electrically connected to an edge computing gateway. The edge computing gateway has built-in data filtering rules and compression algorithms. The output end of the edge computing gateway is electrically connected to a cloud computing platform. The cloud computing platform integrates the Spark real-time analysis engine and the TensorFlow prediction model. The output end of the cloud computing platform is electrically connected to a visual interactive interface. The visual interactive interface supports rendering of three-dimensional garbage distribution situation maps and multi-dimensional data capture and analysis using WebGL technology. Discrete monitoring points are mapped to a unified geographic grid to form a three-dimensional data matrix from time to space to category. Based on a distributed computing framework, the system clusters and detects anomalies in massive data, identifying garbage hotspots and classified distribution patterns. The ST-DBSCAN algorithm, combined with spatiotemporal constraints, effectively improves clustering accuracy, while the sliding window mechanism is used to dynamically screen abnormal data. The analysis results are presented in the form of heat maps and radar maps, intuitively displaying the regional garbage distribution status. At the same time, the system supports historical data backtracking and trend analysis to help management departments understand the long-term patterns of garbage generation and provide data support for the formulation of plans.

[0020] The cloud computing platform uses Kubernetes containers to orchestrate and manage analytical microservices, and the cloud computing platform integrates a Redis in-memory database. The system adopts an edge-cloud collaborative computing architecture. Lightweight models are deployed on edge nodes to filter and compress sensor data in real time, and only key information is uploaded to the cloud, significantly reducing communication overhead. The cloud platform relies on distributed computing resources to perform large-scale data analysis and model training to ensure processing efficiency. Kubernetes container orchestration technology dynamically manages computing resources to adapt to different load requirements. While ensuring low-latency response, this architecture can support complex prediction and optimization tasks and is suitable for large-scale waste management scenarios in smart cities. Edge computing can control data processing delays to milliseconds, while cloud analysis ensures continuous optimization of prediction models and high availability of the system.

[0021] The visual interactive interface includes a GIS map display layer, which overlays and displays the distribution of garbage collection points, the real-time location of removal vehicles, and the status of the landfill. The output end of the GIS map display layer is electrically connected to a data analysis panel, which is electrically connected to an early warning management module. The data analysis panel is used to provide a month-on-month growth rate, a pie chart of classification proportions, and a bar chart of recycling economic value. The early warning management module automatically triggers a red alarm when it detects that the amount of garbage in a region exceeds the carrying capacity threshold. The output end of the early warning management module is electrically connected to a decision support unit, which generates a resource allocation proposal based on the prediction results. The resource allocation proposal includes a vehicle dispatch route and a temporary collection point setting plan. GIS maps dynamically display garbage distribution, the location of collection vehicles, and the status of treatment facilities. WebGL technology enables efficient rendering and supports real-time display of data points. The interactive analysis panel provides multi-dimensional data capture functions, such as classification percentage, regional comparison, and economic value assessment. When the amount of garbage exceeds the preset threshold, the early warning module automatically triggers an alarm and generates an optimized scheduling plan, such as adjusting the collection route or adding temporary collection points. The decision support engine combines prediction results and real-time data to provide managers with scientific resource allocation suggestions, significantly improving garbage management efficiency. The system supports mobile access, ensuring that managers can monitor garbage status anytime, anywhere and respond quickly.

[0022] This application uses a multi-source data collection system to obtain real-time domestic waste data, including multi-dimensional data sources such as smart trash can sensor data, garbage collection truck GPS tracks, and community reporting information, to form a monitoring network covering the entire process of garbage generation, collection, and transportation. The system uses a standardized data processing process to clean and integrate the original information, establish a three-dimensional data matrix containing garbage categories, spatial locations, and time dimensions, and use a distributed computing framework to achieve efficient data analysis and visualization. At the technical implementation level, it systematically integrates edge computing and cloud computing architecture. The edge nodes are responsible for real-time data preprocessing and compression transmission, and the cloud platform relies on big data processing tools such as Spark and TensorFlow for in-depth analysis and model training. At the same time, Kubernetes container technology is used to achieve elastic resource adjustment. The core prediction module is built based on the LSTM neural network. It combines external factors such as meteorology and population mobility to accurately predict the changing trend of garbage volume through time series modeling, and uses an online learning mechanism to continuously optimize model performance. The visual interactive interface integrates GIS maps and WebGL three-dimensional rendering technology, dynamically displaying key information such as garbage distribution heat maps and garbage collection vehicle trajectories, and is equipped with intelligent early warning and decision support functions. When an abnormal situation is detected, it can automatically generate the optimal scheduling plan. The entire system adopts a modular design, supports flexible expansion of new data sources and analysis algorithms, and ensures long-term stability and accuracy through continuous iterative optimization. It provides reliable technical support for the intelligent management of urban domestic waste, effectively improves the efficiency of garbage collection and transportation and the scientific nature of management decisions, and is of great significance to promoting the construction of smart cities and sustainable development.

[0023] The above are all preferred embodiments of the present application, and are not intended to limit the scope of protection of the present application. Therefore, any equivalent changes made based on the structure, shape, and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A method for monitoring the status of regional domestic waste based on big data, characterized by: The following steps are involved: Step 1: Real-time acquisition of domestic waste data in a designated area through a multi-source data collection system; Step 2: Standardize and clean the raw data to create a three-dimensional data matrix containing garbage category identification, spatial location coordinates, and timestamps; Step 3: Use a distributed computing framework to perform cluster analysis on the pre-processed data, generate a regional garbage status heat map and category distribution radar map, and complete the data integration processing; Step 4: Use the time series prediction model combined with external influencing factors to predict future trends in garbage volume.

2. The method for monitoring regional domestic waste status based on big data according to claim 1, characterized in that: The multi-source data acquisition system includes an intelligent trash can sensor data unit, the output end of the intelligent trash can sensor data unit is electrically connected to a garbage collection truck GPS trajectory data unit, the output end of the garbage collection truck GPS trajectory data unit is electrically connected to community reporting data and a government database, and the time series prediction model includes a historical garbage volume data model, a temperature model, a precipitation model and a population flow thermal model.

3. The method for monitoring regional domestic waste status based on big data according to claim 2, characterized in that: The smart trash can sensor data unit is equipped with a pressure sensor. The smart trash can sensor data unit collects data by monitoring the changes in the trash can load in real time through the pressure sensor and establishing a weight-to-capacity correspondence curve. The smart trash can sensor data unit has a built-in infrared spectrum recognition device. The smart trash can sensor data unit collects data by using the infrared spectrum recognition device to classify the materials of the input garbage. The infrared spectrum recognition device is used to distinguish recyclables, kitchen waste and hazardous waste. The smart trash can sensor data unit is integrated with an ultrasonic ranging module. The smart trash can sensor data unit collects data by detecting the garbage filling height through the ultrasonic ranging module. When the ultrasonic ranging module detects that the garbage filling height reaches a preset threshold, it triggers a full load alarm signal. The smart trash can sensor data unit has a built-in NB-IoT communication module. The NB-IoT communication module encapsulates and transmits data in JSON format. The transmission data of the NB-IoT communication module includes device ID, latitude and longitude coordinates and data collection time.

4. The method for monitoring regional domestic waste status based on big data according to claim 1, characterized in that: In step three, more specifically, a data quality assessment system is established, missing data is interpolated and completed using the KNN algorithm, abnormal data is screened and eliminated using the 3σ principle, discrete data points are mapped to 1km×1km geographic units through spatial gridding, and the MapReduce framework of the Hadoop ecosystem is used to implement parallel processing of multi-source heterogeneous data and generate a standardized HDFS storage format.

5. The method for monitoring regional domestic waste status based on big data according to claim 1 is characterized by: The time series prediction model is internally integrated with a prediction and analysis module, which is used to construct an LSTM neural network model. The prediction and analysis module includes an input layer, a hidden layer, and an output layer. The input layer contains 72 hours of historical garbage volume time series data, temperature and humidity data obtained from the Meteorological Bureau API, and population density data analyzed from mobile phone signaling. The hidden layer uses a three-layer GRU unit with a Dropout rate of 0.

2. The output layer generates predicted values ​​for garbage volume in different time periods for the next 24 hours. The prediction and analysis module is equipped with a backpropagation algorithm, which is used to update the model weight parameters.

6. The method for monitoring regional domestic waste status based on big data according to claim 1, characterized in that: The multi-source data acquisition system also includes a distributed data acquisition terminal cluster, the output end of the distributed data acquisition terminal cluster is electrically connected to an edge computing gateway, the edge computing gateway has built-in data filtering rules and compression algorithms, the output end of the edge computing gateway is electrically connected to a cloud computing platform, the cloud computing platform has a Spark real-time analysis engine and a TensorFlow prediction model integrated into it, the output end of the cloud computing platform is electrically connected to a visual interactive interface, and the visual interactive interface supports rendering of a three-dimensional garbage distribution situation map and multi-dimensional data capture and analysis through WebGL technology.

7. The method for monitoring regional domestic waste status based on big data according to claim 6, characterized in that: The cloud computing platform uses Kubernetes containers to orchestrate and manage analysis microservices, and the cloud computing platform is internally integrated with a Redis in-memory database.

8. The method for monitoring regional domestic waste status based on big data according to claim 6, characterized in that: The visual interactive interface includes a GIS map display layer, which superimposes and displays the distribution of garbage collection points, the real-time location of garbage collection vehicles and the status of the landfill. The output end of the GIS map display layer is electrically connected to a data analysis panel, and the output end of the data analysis panel is electrically connected to an early warning management module. The data analysis panel is used to provide a month-on-month growth rate, a category proportion pie chart and a recycling economic value bar chart. The early warning management module automatically triggers a red alarm when it detects that the amount of garbage in the area exceeds the carrying capacity threshold. The output end of the early warning management module is electrically connected to a decision support unit. The decision support unit generates a resource allocation proposal based on the prediction results. The resource allocation proposal includes a vehicle scheduling route and a temporary collection point setting plan.

Images