Smart park real-time data acquisition and processing method and system based on edge computing

Abstract

The invention discloses a smart park real-time data acquisition processing method and system based on edge computing, relates to the technical field of data acquisition, and aims to solve the problem of different display effects caused by a unified processing method for park images at different positions. Setting an acquisition force scoring mechanism to carry out acquisition force scoring on different acquisition devices by utilizing fuzzy reasoning, and carrying out regional division on the park according to acquisition force scores and point location information; detecting the detection area and the fault number of the acquisition equipment in each divided region in real time, calculating the regional real-time coverage rate of each divided region by using a polynomial regression algorithm, classifying each divided region, obtaining a regional image of a low-real-time coverage region, marking the regional image, and acquiring the space dust amount and humidity; and calculating a humidity transfer rate, and judging whether a region image is processed by adopting an enhanced edge calculation algorithm in a corresponding region through a geometric averaging method.

Description

Technical Field

[0001] The present invention relates to the technical field of data acquisition, and more specifically, to a real-time data acquisition and processing method and system for intelligent parks based on edge computing. Background Art

[0002] Data acquisition technology refers to the process of automatically or manually collecting, recording, and processing data. The application of data acquisition technology in real-time data acquisition in intelligent parks can improve the accuracy of data acquisition in intelligent parks. By combining edge computing to process the acquired images, the management resources of the park can be reduced.

[0003] The prior art has the following deficiencies: In the past, when collecting data for intelligent parks, by installing acquisition devices at different locations in the park and obtaining the data information of each location in the park in real-time, without distinguishing the data acquisition locations in the park, uniformly obtaining real-time images for processing, there is a problem that the display effects of the processed images are different, resulting in a decrease in the management efficiency of intelligent parks. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a real-time data acquisition and processing method for an intelligent park based on edge computing. By analyzing the acquisition capabilities of acquisition devices at different locations in the park, the park is divided into regions, and according to the sensor states and environmental differences in different regions, image processing methods for different divided regions are selected to solve the problems proposed in the above background art.

[0005] To achieve the above object, the present invention provides the following technical solutions: A real-time data acquisition and processing method for an intelligent park based on edge computing, comprising the following steps: Step S1: Obtain the attribute data and point location information of each acquisition device in the park, set an acquisition force scoring mechanism, and use the attribute data of the acquisition devices to score the acquisition forces of different acquisition devices. Divide the park into regions according to the acquisition force scores and point location information of each acquisition device; Step S2: Real-time detect the acquisition delay and the number of faults of the acquisition devices in each divided region, use the polynomial regression algorithm to calculate the real-time coverage rate of each divided region, and classify each divided region as a high real-time coverage region or a low real-time coverage region according to the real-time coverage rate of the region; Step S3: Mark the low real-time coverage regions, obtain the regional images of the marked regions, detect the spatial dust amount in each marked region, set the analysis time to collect the humidity in each marked region and calculate the humidity transfer rate; Step S4: Comprehensively use the geometric mean method based on the spatial dust amount and humidity transfer rate of each regional image to determine whether to use an enhanced edge computing algorithm to process the regional image.

[0006] In a preferred embodiment, in step S1, the attribute data and point location information of each acquisition device are called in the local file system of the park. The point location information is the coordinates of the acquisition device in the park, and the attribute data is the acquisition area and working duration of each acquisition device.

[0007] In a preferred embodiment, in step S1, the acquisition force scores of different acquisition devices are determined using fuzzy inference by integrating the acquisition area and working duration of each acquisition device. The specific steps are as follows: Input processing: Use the Z-score standardization algorithm to process the acquisition area and working duration of each acquisition device; Input definition: Define the results after standardizing the acquisition area and the results after standardizing the working duration of each acquisition device as input variables, and divide them into different fuzzy sets; Output definition: Define the acquisition force score of each acquisition device as the output variable; Formulate rules: Formulate fuzzy rules to describe the influence of different inputs on the output; Perform fuzzy inference: Set the acquisition force score of the acquisition device as a high score or a low score according to the output result.

[0008] In a preferred embodiment, in step S1, the park is divided into regions according to the acquisition force scores of each acquisition device and the point location information and the set division rules as follows: Set the division balance threshold, and randomly assign n adjacent high-score acquisition devices and the corresponding number of low-score acquisition devices to the same division region until the entire park is covered, and use the positioning information of each acquisition device to determine the site scope of each division region.

[0009] In a preferred embodiment, in step S2, the time for different types of acquisition devices to collect data until it is transmitted to the host computer, and the number of faults are obtained by counting after fault determination of the acquisition devices in each division region. The specific steps are as follows: In each division region, the acquisition device stores the real-time uploaded data in the temporary database. When performing fault determination, detect the occupied space of the stored data in the temporary database. When the occupied space of the stored data stops increasing, record the current time point to obtain the first time point. When the occupied space of the stored data starts to increase, record the current time point to obtain the second time point. Subtract the first time point from the second time point to obtain the fault time. When the fault time of the acquisition device exceeds the preset fault threshold, it is determined that the acquisition device has a fault, and a count is performed. The fault counts of all acquisition devices in each division region are statistically counted as the number of faults in the corresponding division region.

[0010] In a preferred embodiment, in step S2, after obtaining the acquisition delay and the number of faults of the acquisition devices in each divided area, the regional real-time coverage rate of each divided area is calculated using the polynomial regression algorithm. The specific steps are as follows: Construct a polynomial regression model with the acquisition delay and the number of faults of the acquisition devices in each divided area as inputs. Its formula is: where L is the calculation result of the polynomial regression model, c is the adjustment parameter, a is the acquisition delay of the acquisition devices in the divided area, b is the number of faults in the divided area, and are the input weights of the two inputs of the preset acquisition delay and the number of faults of the acquisition devices respectively. The calculation result of the polynomial regression model is used as the regional real-time coverage rate of each divided area; Calculate the average value of the regional real-time coverage rates of all divided areas as the regional coverage threshold. Classify the divided areas with real-time coverage rates exceeding the regional coverage threshold as high real-time coverage areas. Otherwise, classify them as low real-time coverage areas.

[0011] In a preferred embodiment, in step S3, select a period of time as the analysis time, and collect the humidity of each marked area during the analysis time; Divide the analysis time into N detection times. Collect the humidity of the marked area during each detection time, select the maximum and minimum values of the humidity during the N detection times and calculate the difference. Use the difference result as the humidity transfer rate of the corresponding marked area.

[0012] In a preferred embodiment, in step S4, process the spatial dust amount and the humidity transfer rate of each regional image using the logarithmic transformation normalization formula: , where x is the spatial dust amount or the humidity transfer rate of each regional image, and y is the result of the logarithmic transformation normalization of the spatial dust amount or the humidity transfer rate of each regional image. Respectively use the results of the logarithmic transformation normalization of the spatial dust amount or the humidity transfer rate of each regional image as influence parameter 1 and influence parameter 2 for whether to use the enhanced edge calculation algorithm; Comprehensively use influence parameter 1 and influence parameter 2 to judge whether to process the current regional image through the enhanced edge calculation algorithm using the geometric mean method. The specific steps are as follows: Use influence parameter 1 and influence parameter 2 as inputs and calculate the geometric mean result through the geometric mean method calculation formula: , where k is influence parameter 1, z is influence parameter 2, and h is the geometric mean result; When the geometric mean result exceeds the preset enhancement threshold, the image of the current area is processed using the enhanced edge calculation algorithm; otherwise, it is processed using the ordinary edge calculation algorithm. Among the area images of all marked areas, a random selection is made for judgment until all images are traversed.

[0013] The intelligent park real-time data acquisition and processing system based on edge computing is used to implement the above-mentioned intelligent park real-time data acquisition and processing method based on edge computing, and includes a data acquisition module, a collection force scoring module, a coverage analysis module, and an image processing module. The data acquisition module is used to obtain the attribute data and point location information of each collection device in the park and transmit them to the collection force scoring module; after receiving the divided areas, it performs real-time detection on the collection delay and the number of faults of the collection devices in each divided area, and transmits the detection results to the coverage analysis module. The collection force scoring module is used to receive the attribute data and point location information of each collection device, score the collection force of different collection devices, divide the park into areas, and send each divided area to the data acquisition module. The coverage analysis module calculates the real-time coverage rate of each divided area through the collection delay and the number of faults of the collection devices in each divided area, classifies different divided areas, selects low real-time coverage areas for marking, acquires the area images of the marked areas, detects the spatial dust amount and humidity of the marked areas, and transmits the area images and detection results of the marked areas to the image processing module. The image processing module determines whether to process the area images of different marked areas using the enhanced edge calculation algorithm according to the detection results.

[0014] Technical effects and advantages of the intelligent park real-time data acquisition and processing method and system based on edge computing of the present invention: The present invention obtains the attribute data and point location information of each collection device in the park, sets a collection force scoring mechanism to score the collection force of different collection devices, divides the park into areas according to the collection force scores and point location information of each collection device, evaluates the collection force to screen out key detection areas, performs real-time detection on the detection area and the number of faults of the collection devices in each divided area, calculates the real-time coverage rate of each divided area, classifies each divided area according to the real-time coverage rate, obtains the area images of low real-time coverage areas to reduce the judgment direction and improve the processing efficiency, collects the spatial dust amount of low real-time coverage areas, sets the analysis time to collect the humidity of each area image and calculates the humidity transfer rate, and determines whether to use the enhanced edge calculation algorithm to process the area image of the corresponding area, so as to accurately locate the data source in the park and improve the data acquisition efficiency. Description of the Drawings

[0015] Figure 1Schematic diagram of the real-time data acquisition and processing method for a smart park based on edge computing according to the present invention.

[0016] Figure 2 Flowchart of the real-time data acquisition and processing system for a smart park based on edge computing according to the present invention. Detailed implementation manners

[0017] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0018] The present invention obtains the attribute data and point information of each acquisition device in the park, sets an acquisition force scoring mechanism to score the acquisition force of different acquisition devices, divides the park according to the acquisition force scores and point information of each acquisition device, and performs real-time detection on the detection area and the number of faults of the acquisition devices in each divided area and calculates the real-time coverage rate of each divided area. Classify each divided area according to the real-time coverage rate, obtain the area image of the low real-time coverage area, collect the spatial dust amount of the low real-time coverage area, set the analysis time to collect the humidity of each area image and calculate the humidity transfer rate, and determine whether to use the enhanced edge computing algorithm to process the area image for the corresponding area, so as to accurately locate the data source of the park and improve the data acquisition efficiency.

[0019] Embodiment 1, a real-time data acquisition and processing method for a smart park based on edge computing, as Figure 1 shown, includes the following steps: Step S1: Obtain the attribute data and point information of each acquisition device in the park, set an acquisition force scoring mechanism and use the attribute data of the acquisition device to score the acquisition force of different acquisition devices, and divide the park according to the acquisition force scores and point information of each acquisition device; Step S2: Perform real-time detection on the acquisition delay and the number of faults of the acquisition devices in each divided area, use the polynomial regression algorithm to calculate the real-time coverage rate of each divided area, and classify each divided area into a high real-time coverage area or a low real-time coverage area according to the real-time coverage rate; Step S3: Mark the low real-time coverage area, obtain the area image of the marked area, detect the spatial dust amount of each marked area, set the analysis time to collect the humidity of each marked area and calculate the humidity transfer rate; Step S4: Use the geometric mean method to comprehensively judge whether to use the enhanced edge computing algorithm to process the area image based on the spatial dust amount and humidity transfer rate of each area image.

[0020] The specific implementation is as follows: In step S1, call the attribute data and point location information of each collection device in the local file system of the park. The point location information is the coordinate of the collection device in the park, and the attribute data is the collection area and working duration of each collection device; Use fuzzy inference to determine the collection force scores of different collection devices based on the collection areas and working durations of each collection device. The specific steps are as follows: Input processing: Use the Z-score standardization algorithm to process the collection areas and working durations of each collection device: , where is the collection area or working duration of the collection device, is the average value of the collection areas or working durations of all collection devices, is the standard deviation of the collection areas or working durations of all collection devices, is the result after standardizing the collection area or working duration of the collection device; Input definition: Define the results after standardizing the collection areas of each collection device and the results after standardizing the working durations as input variables and divide them into different fuzzy sets. For example, correspond "Big" and "Small" to the results after standardizing the collection area, and correspond "Long" and "Short" to the results after standardizing the working duration; Output definition: Define the collection force scores of each collection device as output variables. For example, correspond "High" and "Low" to the collection force scores of each collection device.

[0021] Formulate rules: Formulate fuzzy rules to describe the influence of different inputs on the output. For example, mark the result after standardizing the collection area as M, mark the result after standardizing the working duration as G, and mark the collection force score of the collection device as P, then the following can be defined Rule1: IF (M is Big) AND (G is Long) THEN (P is High) Rule2: IF (M is Small) AND (G is Short) THEN (P is Low) Perform fuzzy inference: Set the collection force score of the collection device as a high score or a low score according to the output result; when the output result is "High", it is determined that the corresponding collection device has a high score; when the output result is "Low", it is determined that the corresponding collection device has a low score.

[0022] It should be noted that the classification categories of fuzzy sets can be adjusted according to the actual situation. In this example, two fuzzy sets are used for illustration. In actual operation, three or more fuzzy sets can be selected for classification. In addition, the accuracy threshold for judging the high and low of the input variables in this example can be set by oneself. For example, when the standardized result of the acquisition area of the acquisition device exceeds the average value of the standardized results of the acquisition areas of all acquisition devices, it is calibrated as "Big". The same applies to the accuracy threshold of the standardized result of the working duration of the acquisition device, and no further analysis will be made here.

[0023] According to the acquisition force scores and point position information of each acquisition device, set the division rules to divide the park into regions as follows: Set the division balance threshold, and randomly divide n adjacent high-scoring acquisition devices and the corresponding number of low-scoring acquisition devices into the same division region until the entire park is covered, and use the positioning information of each acquisition device to determine the site scope of each division region.

[0024] It should be noted that the number n of adjacent high-scoring acquisition devices and the corresponding number of low-scoring acquisition devices in each division region is not unique. For example, n can be set to 4, etc., and no further elaboration will be made here.

[0025] In step S2, the acquisition delay of the acquisition device is the time from when the acquisition device of different models in the model database acquires data to when it is transmitted to the upper computer. The number of faults is obtained by counting after fault determination of the acquisition devices in each division region. The specific steps are as follows: In each division region, the acquisition device stores the real-time uploaded data in the temporary database. When performing fault determination, detect the occupied space of the stored data in the temporary database. When the occupied space of the stored data stops increasing, record the current time point to obtain the first time point. When the occupied space of the stored data starts to increase, record the current time point to obtain the second time point. Subtract the second time point from the first time point to obtain the fault time. When the fault time of the acquisition device exceeds the preset fault threshold, it is determined that the acquisition device has a fault, and a count is made. The fault counts of all acquisition devices in each division region are statistically counted as the number of faults in the corresponding division region.

[0026] It should be noted that when obtaining the first time point and the second time point, the acquisition order needs to be followed. The acquisition mechanism of the second time point will be triggered only after the first time point is obtained and recorded. The above-mentioned fault threshold is set by professionals in this field, and no further analysis will be made here.

[0027] The longer the acquisition delay of the acquisition device in the division region, the lower the real-time detection efficiency of the corresponding division region, and the lower the regional real-time coverage rate of the division region; the more the number of faults of the acquisition device in the division region, the lower the real-time detection efficiency of the corresponding division region, and the lower the regional real-time coverage rate of the division region; After obtaining the acquisition delay and the number of faults of the acquisition devices in each divided area, use the polynomial regression algorithm to calculate the real-time coverage rate of each divided area. The specific steps are as follows: Construct a polynomial regression model with the acquisition delay and the number of faults of the acquisition devices in each divided area as inputs. Its formula is: where L is the calculation result of the polynomial regression model, c is the adjustment parameter, a is the acquisition delay of the acquisition device in the divided area, b is the number of faults in the divided area, and are the input weights of the two inputs of the acquisition delay and the number of faults of the acquisition device respectively. Take the calculation result of the polynomial regression model as the real-time coverage rate of each divided area. It should be noted that since the higher the acquisition delay or the failure rate of the acquisition device in the divided area, the lower the real-time coverage rate of the corresponding divided area, the input weights of the two inputs are negative. For example, set and to -0.4 and -0.6 respectively. The adjustment parameter is used to adjust the real-time coverage rate of the divided area to between 0 and 1.

[0028] Calculate the average value of the real-time coverage rates of all divided areas as the area coverage threshold. Classify the divided areas with real-time coverage rates exceeding the area coverage threshold as high real-time coverage areas, and classify the divided areas with real-time coverage rates lower than the area coverage threshold as low real-time coverage areas.

[0029] In step S3, for the low real-time coverage areas, key monitoring is required. Mark the low real-time coverage areas, obtain the area images of the marked areas and the regions through the acquisition devices, receive the spatial dust amounts of different marked areas transmitted by the dust sensors, select a period of time as the analysis time, and collect the humidity of each marked area during the analysis time; Divide the analysis time into N detection times, collect the humidity of the marked areas during each detection time, select the maximum and minimum values of the humidity during the N detection times and take the difference. Take the difference result as the humidity transfer rate of the corresponding marked area; It should be noted that the dust sensor is a device for detecting and measuring the concentration of dust particles in the air. In this example, it is used to collect the spatial dust amounts of the marked areas. When the spatial dust amount is high, the area image acquisition of the corresponding marked area is more blurred, the humidity transfer rate is greater, and the images of the marked area collected at different detection times are more different. The time interval of the detection time can be set according to the actual situation. For example, set the time interval of the detection time to 3 seconds. When calculating the humidity transfer rate of the marked area, collect the humidity of the marked area once during the time interval of each detection time.

[0030] In step S4, the spatial dust amount and humidity transfer rate of each regional image are normalized, and the logarithmic transformation normalization formula is used for processing: , where x is the spatial dust amount or humidity transfer rate of each regional image, and y is the result of the logarithmic transformation normalization of the spatial dust amount or humidity transfer rate of each regional image. The results of the logarithmic transformation normalization of the spatial dust amount or humidity transfer rate of each regional image are used as influence parameter 1 and influence parameter 2 for whether to use the enhanced edge calculation algorithm; Based on influence parameter 1 and influence parameter 2, the geometric mean method is used to determine whether to process the current regional image through the enhanced edge calculation algorithm. The specific steps are as follows: Use influence parameter 1 and influence parameter 2 as inputs and calculate the geometric mean result through the geometric mean method calculation formula: , where k is influence parameter 1, z is influence parameter 2, and h is the geometric mean result. Compare the geometric mean result with the preset enhancement threshold to determine whether to process the current regional image through the enhanced edge calculation algorithm; When the geometric mean result exceeds the enhancement threshold, the current regional image is processed using the enhanced edge calculation algorithm; otherwise, the ordinary edge calculation algorithm is used for processing. Randomly select and judge the regional images of all marked areas until all images are traversed.

[0031] It should be noted that the ordinary edge calculation algorithm is an image processing method based on the prior art. The enhanced edge calculation algorithm removes image noise and redundant image data on the basis of the ordinary edge calculation algorithm. The specific processing steps are carried out by professionals in the field.

[0032] Embodiment 2, a real-time data acquisition and processing system for an intelligent park based on edge computing, as Figure 2 shown, includes a data acquisition module, a collection force scoring module, a coverage analysis module, and an image processing module; The data acquisition module is used to obtain the attribute data and point location information of each acquisition device in the park and transmit them to the collection force scoring module; after receiving the divided areas, it performs real-time detection on the acquisition delay and the number of faults of the acquisition devices in each divided area, and transmits the detection results to the coverage analysis module; The collection force scoring module is used to receive the attribute data and point location information of each acquisition device, score the collection force of different acquisition devices, divide the park into areas, and send each divided area to the data acquisition module; The coverage analysis module calculates the real-time coverage rate of each divided area through the acquisition delay and the number of faults of the acquisition devices in each divided area, classifies different divided areas, selects low real-time coverage areas for marking, acquires the area images of the marked areas, detects the spatial dust amount and humidity of the marked areas, and transmits the area images and detection results of the marked areas to the image processing module; The image processing module determines whether to process the area images of different marked areas using the enhanced edge calculation algorithm according to the detection results.

[0033] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product.

[0034] Those of ordinary skill in the art can realize that the modules and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and the inventive constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0035] In addition, the functional modules in each embodiment of this application can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0036] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0037] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A real-time data collection and processing method for a smart park based on edge computing, characterized in that: The following steps are included: Step S1: Acquire the attribute data and location information of each collection device in the park, set up a collection power scoring mechanism, and use the attribute data of the collection device to score the collection power of different collection devices, and divide the park into regions according to the collection power score and location information of each collection device; Step S2: Perform real-time detection on the acquisition delay and the number of faults of the acquisition equipment in each divided area, calculate the regional real-time coverage rate of each divided area by using a polynomial regression algorithm, and classify each divided area into a high real-time coverage area or a low real-time coverage area according to the regional real-time coverage rate; Step S3: Mark the low real-time coverage area, obtain the regional image of the marked area, detect the amount of dust in the space of each marked area, set the analysis time to collect the humidity of each marked area and calculate the humidity transfer rate; Step S4: Comprehensively analyze the amount of dust in the space and the humidity transfer rate of each regional image and use the geometric mean method to determine whether to use the enhanced edge computing algorithm to process the regional image.

2. The method for real-time data collection and processing in a smart park based on edge computing according to claim 1 is characterized in that: In step S1, the attribute data and point information of each collection device are called in the local file system of the park. The point information is the coordinates of the collection device in the park, and the attribute data is the collection area and working time of each collection device.

3. The real-time data collection and processing method for a smart park based on edge computing according to claim 2 is characterized in that: In step S1, the collection area and working time of each collection device are comprehensively considered and fuzzy reasoning is used to determine the collection force scores of different collection devices. The specific steps are as follows: Input processing: Use the Z-score standardization algorithm to process the collection area and working time of each collection device; Input definition: The standardized results of the collection area and the standardized results of the working hours of each collection device are defined as input variables, and they are divided into different fuzzy sets; Output definition: The collection force score of each collection device is defined as the output variable; Formulate rules: Formulate fuzzy rules to describe the impact of different inputs on outputs; Perform fuzzy reasoning: set the collection power score of the collection device to a high score or a low score according to the output result.

4. The method for real-time data collection and processing in a smart park based on edge computing according to claim 3 is characterized in that: In step S1, the park is divided into regions according to the collection power score of each collection device and the point information setting division rules as follows: A partition balance threshold is set, and n adjacent high-scoring collection devices and a corresponding number of low-scoring collection devices are randomly assigned to the same partition area until the entire park is covered. The site scope of each partition area is determined using the positioning information of each collection device.

5. The method for real-time data collection and processing of a smart park based on edge computing according to claim 1 is characterized in that: In step S2, the time from collecting data for different types of collection devices to transmitting data to the host computer, and the number of faults are obtained by counting and judging the faults of the collection devices in each divided area. The specific steps are as follows: In each divided area, the collection device stores the real-time uploaded data into a temporary database. When making a fault judgment, the space occupied by the stored data in the temporary database is detected. When the space occupied by the stored data stops increasing, the current time point is recorded to obtain the first time point. When the space occupied by the stored data starts to increase, the current time point is recorded to obtain the second time point. The second time point is subtracted from the first time point to obtain the fault time. When the fault time of the collection device exceeds the preset fault threshold, the collection device is judged to be faulty, and a count is performed. The fault counts of all the collection devices in each divided area are counted as the number of faults in the corresponding divided area.

6. The method for real-time data collection and processing in a smart park based on edge computing according to claim 5 is characterized in that: In step S2, after obtaining the collection delay and the number of failures of the collection equipment in each divided area, the real-time coverage rate of each divided area is calculated using a polynomial regression algorithm. The specific steps are as follows: The acquisition delay and the number of failures of the acquisition equipment in each divided area are used as input to construct a polynomial regression model, and its formula is: Among them, L is the calculation result of the polynomial regression model, c is the adjustment parameter, a is the acquisition delay of the acquisition equipment in the divided area, b is the number of faults in the divided area, and The input weights of the two inputs, namely, the acquisition delay of the preset acquisition device and the number of faults, are used as the regional real-time coverage rate of each divided area. The average of the regional real-time coverage rates of all divided areas is calculated as the regional coverage threshold, and the divided areas whose real-time coverage rates exceed the regional coverage threshold are classified as high real-time coverage areas, otherwise, they are classified as low real-time coverage areas.

7. The method for real-time data collection and processing in a smart park based on edge computing according to claim 1 is characterized in that: In step S3, a period of time is selected as the analysis time, and the humidity of each marked area is collected during the analysis time; The analysis time is divided into N detection times. The humidity of the marked area is collected in each detection time. The maximum and minimum humidity values ​​in the N detection times are selected for subtraction, and the subtraction result is used as the humidity transfer rate of the corresponding marked area.

8. The method for real-time data collection and processing in a smart park based on edge computing according to claim 7 is characterized in that: In step S4, the spatial dust volume and humidity transfer rate of each area image are processed using the logarithmic transformation normalization formula: , where x is the amount of dust or the humidity transfer rate of each regional image, and y is the result of the normalization of the amount of dust or the humidity transfer rate of each regional image after logarithmic transformation. The results of the normalization of the amount of dust or the humidity transfer rate of each regional image after logarithmic transformation are used as the influencing parameters 1 and 2 of whether to adopt the enhanced edge computing algorithm; The geometric mean method is used to comprehensively analyze the influencing parameters 1 and 2 to determine whether to process the current area image through the enhanced edge computing algorithm. The specific steps are as follows: The geometric mean result is calculated by using the geometric mean method formula with influencing parameters 1 and 2 as input: , where k is the influencing parameter 1, z is the influencing parameter 2, and h is the geometric mean result; When the geometric mean result exceeds the preset enhancement threshold, the current area image is processed using the enhanced edge computing algorithm, otherwise, it is processed using the ordinary edge computing algorithm; random selection is made from the area images of all marked areas for judgment until all images are traversed.

9. A real-time data collection and processing system for a smart park based on edge computing, based on the real-time data collection and processing method for a smart park based on edge computing according to any one of claims 1 to 8, characterized in that: It includes data acquisition module, acquisition force scoring module, coverage analysis module and image processing module; The data collection module is used to obtain the attribute data and point information of each collection device in the park and transmit it to the collection force scoring module; after receiving the divided area, the collection delay and the number of faults of the collection equipment in each divided area are detected in real time, and the detection results are transmitted to the coverage analysis module; The collection power scoring module is used to receive the attribute data and point information of each collection device, score the collection power of different collection devices and divide the park into regions, and send each divided region to the data collection module; The coverage analysis module calculates the real-time coverage rate of each divided area through the acquisition delay and the number of faults of the acquisition equipment in each divided area, classifies the different divided areas, selects the low real-time coverage area for marking, collects the regional image of the marked area, detects the amount of dust and humidity in the space of the marked area, and transmits the regional image of the marked area and the detection result to the image processing module; The image processing module determines whether to use the enhanced edge computing algorithm to process the regional images of different marked areas according to the detection results.

Images