Park inspection system based on internet of things

Abstract

The application provides a kind of based on Internet of Things's garden inspection system, it is related to the technical field of garden patrol.The system includes data acquisition module, household risk assessment module, inspection system generation module, inspection execution module, data feedback module and household management module.Through the collection of real-time data, risk assessment model is constructed to judge risk level, generate patrol plan and dynamically adjust, execute patrol and record abnormality, abnormality analysis and notify the household.The application can identify potential safety hazards in time by collecting environmental monitoring data and household activity data, and ensure that preventive measures are taken before problems occur.By analyzing real-time data, accurate risk level is generated to ensure that high-risk areas are given priority attention, effectively reducing the probability of safety incidents.Optimizing the patrol plan, combined with risk level and scheduling algorithm, ensures that the patrol work is targeted and efficient.

Description

Technical Field

[0001] This invention relates to the field of park inspection technology, and in particular to a park inspection system based on the Internet of Things. Background Technology

[0002] The Internet of Things (IoT)-based park inspection system is a new type of system that utilizes advanced technologies such as IoT, artificial intelligence, and data analytics to achieve intelligent management and inspection of parks. Its main purpose is to improve the efficiency of park security management and residents' sense of security through real-time data collection, analysis, and feedback.

[0003] While existing park inspection systems have achieved some success in management, several significant shortcomings remain. Traditional systems typically rely on manual patrols at fixed frequencies, failing to effectively address dynamically changing environments. This model makes patrol effectiveness susceptible to various factors, such as weather, the inspector's subjective judgment, and the design of patrol routes. When emergencies or safety hazards occur within the park, management personnel often cannot respond promptly, limiting patrol effectiveness. Furthermore, traditional inspection models typically cannot dynamically adjust patrol plans based on real-time conditions, leading to wasted resources. Summary of the Invention

[0004] This invention provides an Internet of Things-based park inspection system to address the shortcomings of existing technologies.

[0005] This invention provides an Internet of Things (IoT) based park inspection system, comprising:

[0006] The data acquisition module is used to collect real-time data from residents within the park through data acquisition equipment. The real-time data includes environmental monitoring data and resident activity data.

[0007] The resident risk assessment module is used to build a risk assessment model based on convolutional neural networks, inputting real-time data and outputting the risk level of residents in the park.

[0008] The inspection system generation module is used to generate inspection plans based on risk levels. The inspection plans include inspection frequency and inspection routes, and the scheduling algorithm is used to optimize the inspection plans.

[0009] The inspection execution module is used to inspect the park according to the inspection plan and record abnormal data.

[0010] The data feedback module performs time series analysis on abnormal data, generates analysis reports, evaluates the effectiveness of inspections based on the analysis reports, and adjusts the inspection plan accordingly.

[0011] The resident management module is used to manage resident information, set permissions, and provide notification functions, and to notify residents in the park of the inspection results.

[0012] According to the present invention, a park inspection system based on the Internet of Things (IoT) includes a data acquisition module comprising an environmental monitoring unit and a resident activity monitoring unit. The environmental monitoring unit uses sensor devices to monitor environmental data of residents within the park in real time, including temperature, humidity, smoke concentration, and noise data. The resident activity monitoring unit uses access control systems and surveillance cameras to record the entry and exit times of residents within the park in real time, and calculates activity frequency data based on this entry and exit time data.

[0013] According to the IoT-based park inspection system provided by the present invention, the process of constructing a risk assessment model based on a convolutional neural network includes:

[0014] Collect historical environmental monitoring data and historical resident activity data within the park, and analyze the resident risk level based on the historical environmental monitoring data and historical resident activity data.

[0015] Historical environmental monitoring data and historical resident activity data are preprocessed to obtain preprocessed data. Preprocessing includes handling missing values, removing outliers, and data standardization.

[0016] A basic model based on a convolutional neural network is constructed. Preprocessed data is used as input and the risk level of residents is used as output. The basic model is trained, and the model parameters that meet the test accuracy are retained to obtain the risk assessment model.

[0017] According to the present invention, a park inspection system based on the Internet of Things (IoT) includes a basic model based on a convolutional neural network (CNN), comprising convolutional layers, pooling layers, fully connected layers, and an output layer. The convolutional layers are used to design the basic CNN network structure, which includes multiple convolutional layers and activation layers. The convolutional layers extract local features from the input feature map through convolution operations.

[0018] Pooling layers are used to reduce the dimensionality of the input feature map through max pooling, thereby reducing computation and preventing overfitting.

[0019] Fully connected layers are used to integrate the extracted features.

[0020] The output layer is used to transform the output of the base model into class probabilities by setting the Softmax activation function.

[0021] According to the present invention, a park inspection system based on the Internet of Things (IoT) includes a risk assessment result processing unit, a patrol route planning unit, and a scheduling algorithm unit. The risk assessment result processing unit receives the risk level output by the resident risk assessment module and converts the risk level into input parameters for the patrol plan. The patrol route planning unit generates patrol routes based on the distribution and risk levels of residents within the park and optimizes the patrol routes using Dijkstra's algorithm.

[0022] The scheduling algorithm unit is used to optimize the inspection plan based on the genetic algorithm.

[0023] According to the present invention, a park inspection system based on the Internet of Things (IoT) includes an inspection execution module comprising an inspection task management unit and an inspection execution and data recording unit. The inspection task management unit receives inspection plans, schedules resources according to the plans, and updates real-time task status. The inspection execution and data recording unit executes inspection tasks according to the plans and records abnormal data during the inspection process in real time, including abnormal environmental data and abnormal activity data.

[0024] According to the present invention, a park inspection system based on the Internet of Things includes a data feedback module comprising an abnormal data processing unit and a time series analysis and anomaly detection unit. The abnormal data processing unit receives abnormal data and processes it to obtain processed data, including missing value handling, data cleaning, and standardization. The time series analysis and anomaly detection unit uses an LSTM model to perform time series analysis on the processed data to identify trends, periodicity, and potential abnormal patterns.

[0025] According to the Internet of Things-based park inspection system provided by the present invention, the data feedback module further includes an effect evaluation and optimization suggestion unit; the effect evaluation and optimization suggestion unit is used to evaluate the inspection effect based on the time series analysis results and generate optimization suggestions for the inspection plan.

[0026] According to the present invention, a park inspection system based on the Internet of Things is provided, wherein the resident management module includes a resident information management unit; the resident information management unit is used to collect and store basic information of residents in the park, including resident's personal information, contact information, address and permission settings, and provides a resident interface for residents to view and update basic information.

[0027] According to the Internet of Things-based park inspection system provided by the present invention, the resident management module further includes a notification and feedback unit; the notification and feedback unit is used to send abnormal data to the target resident.

[0028] Compared with the prior art, the beneficial effects of the present invention are at least as follows:

[0029] This invention provides an IoT-based park inspection system that constructs a risk assessment model based on convolutional neural networks. It takes real-time data as input and outputs resident risk levels. Based on the model's risk assessment, it identifies potential safety hazards, enabling timely preventative measures to be taken before safety issues occur. Simultaneously, it rationally allocates patrol resources based on the risk assessment model's output, prioritizing high-risk areas and taking preventative measures before risks occur in these areas, thus reducing the probability of safety incidents. By combining risk levels with scheduling algorithms to optimize patrol plans, the system makes patrols more targeted and efficient. Furthermore, optimizing patrol frequency and routes improves patrol efficiency, saving manpower and time. By recording abnormal data in real time and responding quickly to and handling anomalies, the system maximizes the park's safety response capabilities. Time-series analysis of abnormal data and the generation of detailed analysis reports help managers evaluate patrol effectiveness, dynamically adjust patrol plans, and continuously optimize the park's safety management strategy. Attached Figure Description

[0030] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0031] Figure 1 This is a schematic diagram of the structure of the Internet of Things-based park inspection system provided in an embodiment of the present invention;

[0032] Figure 2 This is a flowchart of the IoT-based park inspection system provided in an embodiment of the present invention;

[0033] Figure 3 This is a schematic diagram of the structure of the inspection system generation module in an embodiment of the present invention. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0035] The following is combined Figures 1-3This invention describes an Internet of Things (IoT)-based campus inspection system.

[0036] Figure 1 This is a schematic diagram of the structure of the Internet of Things-based park inspection system provided in an embodiment of the present invention; Figure 2 This is a flowchart of the Internet of Things-based park inspection system provided in an embodiment of the present invention.

[0037] like Figures 1-2 As shown, the IoT-based park inspection system provided in this embodiment of the invention includes a data acquisition module, a resident risk assessment module, an inspection system generation module, an inspection execution module, a data feedback module, and a resident management module.

[0038] The data acquisition module is used to collect real-time data from residents within the park through data acquisition equipment. The real-time data includes environmental monitoring data and resident activity data.

[0039] The data acquisition module includes an environmental monitoring unit and a resident activity monitoring unit. The environmental monitoring unit uses sensor devices to monitor environmental data from residents within the park in real time. This data includes temperature, humidity, smoke concentration, and noise levels. The resident activity monitoring unit uses access control and surveillance cameras to record the entry and exit times of residents within the park in real time, and calculates activity frequency data based on this data. The formula is as follows:

[0040]

[0041] In the formula, S represents the activity frequency, and C frequency H represents the number of activities within a preset time period. time This indicates the total duration of the preset time period.

[0042] In this embodiment, the data acquisition module collects real-time data of residents in the park, including environmental monitoring data and resident activity data. The patrol plan can be flexibly adjusted based on the collected real-time data of residents in the park, and the information obtained during the park inspection is more scientific and comprehensive.

[0043] The resident risk assessment module is used to build a risk assessment model based on convolutional neural networks, inputting real-time data and outputting the risk level of residents in the park.

[0044] The process of building a risk assessment model based on convolutional neural networks includes:

[0045] Collect historical environmental monitoring data and historical resident activity data within the park, and analyze the resident risk level based on the historical environmental monitoring data and historical resident activity data.

[0046] Historical environmental monitoring data and historical resident activity data are preprocessed to obtain preprocessed data. Preprocessing includes handling missing values, removing outliers, and data standardization. The formula for data standardization is expressed as:

[0047]

[0048] In the formula, X represents the original data, μ represents the mean, and σ represents the standard deviation.

[0049] We construct a basic model based on convolutional neural networks (CNNs). This basic model includes convolutional layers, pooling layers, fully connected layers, and an output layer. Convolutional layers are used to design the network structure of the basic CNN. The network structure includes multiple convolutional layers and activation layers. Convolutional layers extract local features through convolution operations. The formula for a convolution operation is expressed as:

[0050]

[0051] In the formula, f represents the input feature map, g represents the convolution kernel, i and j represent the position coordinates, i.e. the position of the pixel in the input feature map, and m and n represent the relative coordinates of the convolution kernel.

[0052] Pooling layers are used to reduce the dimensionality of the input feature map through max pooling, thereby reducing computational cost and preventing overfitting. The formula for max pooling is expressed as:

[0053]

[0054] In the formula, S represents the output feature map of the pooling layer.

[0055] Fully connected layers are used to integrate the extracted features.

[0056] The output layer is used to transform the output of the base model into class probabilities by setting the Softmax activation function. The formula for the Softmax activation function is as follows:

[0057]

[0058] In the formula, P(y i z represents the predicted probability of the i-th class. i This represents the raw score of the input data in category i. This represents the summation of the index values ​​for all categories j.

[0059] Using preprocessed data as input and resident risk levels as output, the basic model is trained, and model parameters that meet the testing accuracy are retained to obtain the risk assessment model.

[0060] In this embodiment, a large amount of historical data is collected, including historical environmental monitoring data and historical resident activity data of residents in the park. The collected historical environmental monitoring data and historical resident activity data are manually labeled with risk levels, and the collected data are preprocessed. Then, the preprocessed historical environmental monitoring data and historical resident activity data are used as input to a risk assessment model based on a convolutional neural network, and the risk level is used as output to train the risk assessment model based on the convolutional neural network. After meeting the preset output accuracy, the model parameters with the highest accuracy are retained as the model parameters.

[0061] Figure 3 This is a schematic diagram of the structure of the inspection system generation module in an embodiment of the present invention.

[0062] like Figure 3 As shown, the inspection system generation module is used to generate inspection plans based on risk levels. The inspection plans include inspection frequency and inspection routes, and the scheduling algorithm is used to optimize the inspection plans.

[0063] The inspection system's generation module includes a risk assessment result processing unit, an inspection route planning unit, and a scheduling algorithm unit. The risk assessment result processing unit receives the risk level output from the resident risk assessment module and converts it into input parameters for the inspection plan. Residents with high-risk levels require more frequent inspections, while those with low-risk levels can have their inspection frequency reduced. Inspection priorities are set based on risk levels to ensure that high-risk areas are inspected first.

[0064] The patrol route planning unit is used to generate patrol routes based on the distribution of residents and risk levels within the park, and the Dijkstra algorithm is used to optimize the patrol routes. The formula is expressed as:

[0065]

[0066] In the formula, d(v) represents the shortest path length to node v, and w(u,v) represents the weight of edge u to v.

[0067] The scheduling algorithm unit is used to optimize the inspection plan based on the genetic algorithm, and the formula is expressed as:

[0068]

[0069] In the formula, F(x) represents the fitness of the current solution, ω i d represents the weight of task i. i (x) represents the patrol distance assigned to task i.

[0070] In this embodiment, the park inspection system needs to obtain a geographical map of the park and a location distribution map of the residents within the park. Based on the density of residents within the park and the risk level of each resident output by the risk assessment model, an inspection plan is formulated, including inspection routes and frequencies. The system increases the number of inspections in areas with higher resident density and for residents with higher risk assessment levels, thereby minimizing the possibility of unexpected anomalies within the park and providing a safer environment for the residents.

[0071] The inspection execution module is used to inspect the park according to the inspection plan and record abnormal data.

[0072] The inspection execution module includes an inspection task management unit, an inspection execution and data recording unit, and a real-time monitoring and feedback analysis unit. The inspection task management unit receives inspection plans, allocates resources according to the plans, and updates real-time task status. The inspection execution and data recording unit executes inspection tasks according to the plans and records abnormal data during the inspection process in real time, including abnormal environmental data and abnormal activity data.

[0073] In this embodiment, historical data from residents within the park can be collected to set safety thresholds for each resident. This includes data on temperature, humidity, and frequency of activity for residents on weekdays and weekends. For special residents, such as those producing or using special gases, the concentration of these gases at their residences on weekdays and weekends also needs to be collected. During inspections, this data is analyzed in real time and compared with preset thresholds. If the data exceeds the threshold range, it is considered abnormal.

[0074] The data feedback module performs time series analysis on abnormal data, generates analysis reports, evaluates the effectiveness of inspections based on the analysis reports, and adjusts the inspection plan accordingly.

[0075] The data feedback module includes an anomaly data processing unit, a time series analysis and anomaly detection unit, and an effectiveness evaluation and optimization suggestion unit. The anomaly data processing unit receives and processes anomaly data, including missing value handling, data cleaning, and standardization. The time series analysis and anomaly detection unit uses an LSTM model to perform time series analysis on the processed data, identifying trends, periodicity, and potential anomaly patterns. The effectiveness evaluation and optimization suggestion unit evaluates the patrol effectiveness based on the time series analysis results and generates optimization suggestions for the patrol plan. The formula for evaluating patrol effectiveness is expressed as:

[0076]

[0077] In the formula, E represents the inspection effectiveness evaluation index, C represents the number of abnormal situations resolved, and T represents the total number of abnormal situations detected during the inspection.

[0078] The resident management module is used to manage resident information, set permissions, and provide notification functions, and to notify residents in the park of the inspection results.

[0079] The resident management module includes a resident information management unit and a notification and feedback unit. The resident information management unit collects and stores basic information about residents within the park, including personal details, contact information, address, and access permissions. It also provides a resident interface for viewing and updating this basic information. The notification and feedback unit sends notifications to the relevant resident if abnormal data is detected.

[0080] Example 1: A small residential complex has 10 households, distributed as follows: Household A, Unit 101, Building 1; Household B, Unit 102, Building 1; Household C, Unit 201, Building 1; Household D, Unit 202, Building 1; Household E, Unit 101, Building 2; Household F, Unit 102, Building 2; Household G, Unit 201, Building 2; Household H, Unit 202, Building 2; Household I, Unit 101, Building 3; Household J, Unit 102, Building 3.

[0081] Real-time data collection from each household within the park yielded the following monitoring results for 10 households: Temperature: 22℃, 23℃, 21℃, 24℃, 22℃, 23℃, 25℃, 26℃, 22℃, 21℃; Humidity: 50%, 55%, 52%, 48%, 51%, 53%, 54%, 49%, 50%, 52%; Smoke Concentration: 0.01ppm, 0.02ppm, 0.01ppm, 0.03ppm, 0.01ppm, 0.02ppm, 0.01ppm, 0.04ppm, 0.01ppm, 0.02ppm; Noise Level: 30dB, 35dB, 32dB, 40dB, 30dB, 33dB, 29dB, 38dB, 31dB, 34dB. The monitored resident activity data are as follows: Resident A's entry and exit times are: 8, 12, 18, 22; frequency: 4 times per day. Resident B's entry and exit times are: 6, 18; frequency: 2 times per day. Resident C's entry and exit times are: 7, 9, 15, 18; frequency: 4 times per day. Resident D's entry and exit times are: 8, 18; frequency: 2 times per day. Resident E's entry and exit times are: 8, 11, 18, 20, 22; frequency: 5 times per day. Resident F's entry and exit times are: 7, 11, 16; frequency: 3 times per day. Resident G's entry and exit times are: 8, 19; frequency: 2 times per day. Resident H's entry and exit times are: 9, 11, 13, 18; frequency: 4 times per day. Resident I's entry and exit times are: 9, 18; frequency: 2 times per day. Resident J's entry and exit times are: 8, 18; frequency: 2 times per day.

[0082] The aforementioned monitored temperature data, humidity data, smoke concentration data, noise data, and activity time and frequency were input into the trained convolutional neural network-based risk assessment model, which outputs the risk levels of residents in the park as follows: Resident A: Risk Level 3; Resident B: Risk Level 2; Resident C: Risk Level 4; Resident D: Risk Level 1; Resident E: Risk Level 5; Resident F: Risk Level 3; Resident G: Risk Level 2; Resident H: Risk Level 4; Resident I: Risk Level 2; Resident J: Risk Level 1.

[0083] Based on the preset risk level classification criteria, the risk levels of the 10 households in the park are divided into high-risk households (risk level 4 and above): households C, E, and H; medium-risk households (risk level 3): households A and F; and low-risk households (risk levels 1 and 2): households B, D, G, I, and J.

[0084] Patrol routes are set based on the risk level of residents. For example, during a single patrol, it is necessary to pass through high-risk residents 3 times, medium-risk residents 2 times, and low-risk residents 1 time. An optimization algorithm is used to generate patrol routes, and a scheduling algorithm is used to further optimize them.

[0085] Set safety thresholds for environmental monitoring data and resident activity data, such as: temperature: 30℃; humidity: 40% to 70%; smoke concentration: 0.05ppm; noise: 50dB.

[0086] During inspections, if environmental monitoring data or resident activity data for a particular household exceeds a preset safety threshold, it is considered an anomaly. Inspectors should thoroughly investigate the cause of the anomaly and promptly notify the resident. For example, if the smoke concentration in resident E is 0.06 ppm, exceeding the threshold, it is considered abnormal. Inspectors should check the smoke alarm, ventilation, and for any open flames or other abnormalities, and contact resident E immediately.

[0087] In summary, this embodiment provides an IoT-based park inspection system. By constructing a risk assessment model based on convolutional neural networks, it takes real-time data as input and outputs resident risk levels. Based on the model's risk assessment, potential safety hazards are identified, enabling timely preventative measures to be taken before safety issues occur. Simultaneously, patrol resources are rationally allocated based on the risk assessment model's output, prioritizing high-risk areas and taking preventative measures before risks occur in these areas, reducing the probability of safety incidents. By combining risk levels with scheduling algorithms to optimize patrol plans, patrol work becomes more targeted and efficient. Furthermore, optimizing patrol frequency and routes improves patrol efficiency, saving manpower and time. By recording abnormal data in real time and responding quickly to and handling anomalies, the system maximizes the park's security response capabilities. Time-series analysis of abnormal data and the generation of detailed analysis reports help managers evaluate patrol effectiveness, dynamically adjust patrol plans, and continuously optimize the park's security management strategy.

[0088] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0089] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A park inspection system based on the Internet of Things, characterized in that, include: The data acquisition module is used to collect real-time data from residents within the park through data acquisition equipment. This real-time data includes environmental monitoring data and resident activity data. The data acquisition module includes an environmental monitoring unit and a resident activity monitoring unit. The environmental monitoring unit is used to monitor environmental data from residents within the park in real-time using sensor equipment. This environmental monitoring data includes temperature data, humidity data, smoke concentration data, and noise data. The resident activity monitoring unit is used to record the entry and exit times of residents within the park in real-time through the access control system and surveillance cameras, and calculate activity frequency data based on the entry and exit time data. The resident risk assessment module is used to build a risk assessment model based on a convolutional neural network. It takes the real-time data as input and outputs the risk level of residents in the park. The inspection system generation module is used to generate an inspection plan based on the risk level. The inspection plan includes the inspection frequency and inspection route, and the inspection plan is optimized by combining a scheduling algorithm. The inspection execution module is used to inspect the park according to the inspection plan and record abnormal data; The inspection execution module includes an inspection task management unit and an inspection execution and data recording unit; the inspection task management unit is used to receive the inspection plan, schedule resources according to the inspection plan, and update the real-time task status. The patrol execution and data recording unit is used to execute patrol tasks according to the patrol plan and record abnormal data in real time during the patrol process. The abnormal data includes abnormal environmental data and abnormal activity data. The inspection execution module collects historical data from residents within the park and sets safety thresholds for each resident. These safety thresholds include monitoring data on temperature, humidity, and frequency of human activity for residents on weekdays and weekends. For special residents who produce or use special gases, the module collects the concentration of the special gases present on weekdays and weekends. During the inspection process, this data is analyzed in real time and compared with preset thresholds. If the data exceeds the threshold range, it is considered abnormal. The data feedback module performs time series analysis on the abnormal data, generates an analysis report, evaluates the patrol effectiveness based on the analysis report, and adjusts the patrol plan based on the patrol effectiveness. The resident management module is used to manage resident information, set permissions, and provide notification functions, and to notify residents in the park of the inspection results.

2. The IoT-based park inspection system according to claim 1, characterized in that, The process of building a risk assessment model based on convolutional neural networks includes: Collect historical environmental monitoring data and historical resident activity data of residents in the park, and analyze the resident risk level based on the historical environmental monitoring data and historical resident activity data; The historical environmental monitoring data and historical resident activity data are preprocessed to obtain preprocessed data; the preprocessing includes handling missing values, removing outliers, and data standardization. A basic model based on a convolutional neural network is constructed. The preprocessed data is used as input and the risk level of the residents is used as output. The basic model is trained, and the model parameters that meet the test accuracy are retained to obtain a risk assessment model.

3. The IoT-based park inspection system according to claim 2, characterized in that, The basic model based on the convolutional neural network includes convolutional layers, pooling layers, fully connected layers, and an output layer; the convolutional layers are used to design the network structure of the convolutional neural network, and the network structure includes multiple convolutional layers and activation layers. The convolutional layers extract local features of the input feature map through convolution operations. The pooling layer is used to reduce the dimensionality of the input feature map through max pooling, thereby reducing computation and preventing overfitting. The fully connected layer is used to integrate the extracted features; The output layer is used to convert the output of the base model into class probabilities by setting the Softmax activation function.

4. The IoT-based park inspection system according to claim 1, characterized in that, The inspection system generation module includes a risk assessment result processing unit, an inspection route planning unit, and a scheduling algorithm unit. The risk assessment result processing unit is used to receive the risk level output by the resident risk assessment module and convert the risk level into input parameters for the inspection plan. The inspection route planning unit is used to generate inspection routes based on the distribution and risk level of residents in the park and to optimize the inspection routes using the Dijkstra algorithm. The scheduling algorithm unit is used to optimize the inspection plan based on a genetic algorithm.

5. The IoT-based park inspection system according to claim 1, characterized in that, The data feedback module includes an abnormal data processing unit and a time series analysis and anomaly detection unit. The abnormal data processing unit receives the abnormal data and processes it to obtain processed data. The processing includes missing value handling, data cleaning, and standardization. The time series analysis and anomaly detection unit uses an LSTM model to perform time series analysis on the processed data to identify trends, periodicity, and potential anomaly patterns.

6. The IoT-based park inspection system according to claim 5, characterized in that, The data feedback module also includes an effect evaluation and optimization suggestion unit; the effect evaluation and optimization suggestion unit is used to evaluate the patrol effect based on the time series analysis results and generate optimization suggestions for the patrol plan.

7. The IoT-based park inspection system according to claim 1, characterized in that, The resident management module includes a resident information management unit; the resident information management unit is used to collect and store basic information of residents in the park, including resident's personal information, contact information, address and permission settings, and provides a resident interface for residents to view and update the basic information.

8. The IoT-based park inspection system according to claim 7, characterized in that, The resident management module also includes a notification and feedback unit; the notification and feedback unit is used to send the abnormal data to the target resident.

Images