Public management space monitoring method and device based on unmanned aerial vehicle image vision algorithm

Abstract

The invention discloses a public management space monitoring method and device based on an unmanned aerial vehicle image vision algorithm, and particularly relates to the technical field of monitoring management.The method comprises the steps that a target park area is shot through an unmanned aerial vehicle and analyzed, and the possible environment problem of the target park area is obtained; and then, according to the possible environmental problems of the target park area, appropriate workers are dispatched to further monitor the environment of the target park area, so that the workers can be managed and dispatched timely and reasonably according to the environmental state in the park; the monitoring effect of workers in the park on the park environment is high, the workers can accurately monitor the environment problem in the park in time, and the influence on the overall environment quality of the park and tourist experience is reduced.

Description

Technical Field

[0001] The present invention relates to the field of monitoring and management technology, and in particular to a method and device for monitoring a public management space based on an unmanned aerial vehicle image vision algorithm. Background Art

[0002] Parks provide people with a place to get in touch with nature, relax, and engage in outdoor activities. They can also beautify the environment and improve its quality. Especially some large parks, relying on their own unique scenery, can attract tourists to visit, which not only beautifies the environment but also brings economic benefits.

[0003] Therefore, for large-scale environmental parks, in order to ensure the cleanliness and beauty of the park environment, staff are generally set up in various areas of the park to monitor environmental problems in the park in a timely manner. Once there is a problem with the environment in a certain area of ​​the park, corresponding measures can be taken in time to ensure the beauty of the park environment. In addition, these staff will also conduct regular inspections and maintenance work, including cleaning up garbage, trimming grass and trees, maintaining facilities, etc., to prevent potential problems from developing into greater hidden dangers. Through such monitoring, management and maintenance measures, not only can a good experience for tourists be guaranteed, but the park’s natural resources and ecological environment can also be effectively protected, promoting the development of sustainable tourism.

[0004] However, among the above methods, this manual reliance is inefficient and cannot timely and reasonably manage and dispatch staff according to the environmental conditions in the park, resulting in low monitoring effect of the park environment by the staff in the park. This may eventually lead to the staff being unable to timely and accurately monitor environmental problems in the park, affecting the overall environmental quality of the park and the visitor experience. Summary of the invention

[0005] The purpose of the present invention is to solve the problem that staff cannot monitor the environment in the park in a timely and accurate manner, which affects the overall environmental quality of the park and the visitor experience, and to propose a public management space monitoring method and device based on drone image vision algorithm.

[0006] In the first aspect of the implementation of the present invention, a public management space monitoring method based on a drone image vision algorithm is first proposed, the method comprising:

[0007] Acquire an image set of a target park area through a drone to obtain a first image set, and perform feature extraction on each image in the first image set to obtain image quality feature data of the image;

[0008] For each image in the first image set, determine a preset quality level of the image according to the image quality feature data, and select images that meet the required preset quality level as the second image set; the preset quality levels include high level, medium level and low level, and the required preset quality level is high level;

[0009] Inputting the images in the second image set into a preset environmental problem type model, and outputting an environmental problem type set for the target park area;

[0010] The staff that can be finally dispatched are determined based on the set of environmental problem types and staff status data to monitor the environment of the target park area.

[0011] Optionally, performing feature extraction on each image in the first image set to obtain image quality feature data of the image includes:

[0012] The image quality characteristic data includes the edge clarity of the image, and the edge clarity is obtained by:

[0013] Calculating the gradient in the horizontal direction and the gradient in the vertical direction of each image in the first image set, and obtaining an edge intensity image according to the gradient in the horizontal direction and the gradient in the vertical direction;

[0014] Calculating the second-order derivative of the edge intensity image to obtain a transformed image;

[0015] The average pixel value of the transformed image is calculated, and the variance of the changed image is calculated according to the average pixel value, and the variance of the changed image is used as the edge clarity of the image.

[0016] Optionally, the image quality feature data also includes key area coverage and illumination uniformity;

[0017] The method for obtaining the key area coverage includes:

[0018] Obtain several key areas of the target park area and determine the important weight value of each key area;

[0019] Obtaining the area of ​​different key regions corresponding to each image in the first image set, and multiplying the area of ​​each key region by the corresponding weight value to obtain the weighted area of ​​the key region;

[0020] The weighted areas of all key areas of each image are added together to obtain the total weighted area, and the total weighted area is divided by the area of ​​the image to obtain the key area coverage of the image;

[0021] The method for obtaining the illumination uniformity includes:

[0022] Each image in the first image set is converted into a grayscale image, and the average value of the grayscale values ​​of all pixels in the grayscale image is calculated as the mean, and the standard deviation of the image is calculated according to the mean, and the reciprocal of the standard deviation is used as the illumination uniformity of the image.

[0023] Optionally, determining a preset quality level of the image according to the image quality characteristic data includes:

[0024] The image quality feature data includes edge clarity, key area coverage and illumination uniformity of the image;

[0025] The edge definition, key area coverage and illumination uniformity of the image are defined as the input items of the fuzzy rules and classified into different fuzzy sets respectively;

[0026] The preset quality level is used as the output variable and divided into fuzzy sets;

[0027] Formulate fuzzy rules to describe the influence of different input variables on output variables;

[0028] Perform fuzzy reasoning based on fuzzy rules and output the preset quality level of the image.

[0029] Optionally, the workers that can be finally dispatched are determined according to the set of environmental problem types and the worker status data, including:

[0030] Determine a preliminary set of dispatchable workers according to the type of environmental problem, and obtain status data of each worker in the preliminary set of dispatchable workers;

[0031] The staff member's status data includes a travel distance coefficient, a workload coefficient and a function evaluation coefficient;

[0032] The method for obtaining the driving distance coefficient is:

[0033] The shortest drivable distance from each staff member to the target environmental area, as well as the average speed of the vehicle and the preset normal driving speed corresponding to the shortest drivable distance are obtained to obtain the staff member's driving distance coefficient; the calculation formula is: HK = ws / yh × ut, where HK is the driving distance coefficient, ws, yh and ut are the average speed of the vehicle, the preset normal driving speed and the shortest drivable distance respectively;

[0034] The method for obtaining the workload coefficient is:

[0035] Obtain the current unfinished task amount, maximum task load, and free time from the last task completion of each staff member, and obtain the staff member's workload coefficient. The calculation formula is: GF = ds × fg / th, where GF is the workload coefficient, th, ds, and fg are the current unfinished task amount, maximum task load, and free time from the last task completion, respectively;

[0036] The method for obtaining the function evaluation coefficient is as follows:

[0037] The number of times each staff member worked and the accuracy of monitoring environmental issues were obtained based on historical data, and the number of times each staff member worked and the accuracy of monitoring environmental issues were weighted and summed to obtain the staff member's function evaluation coefficient;

[0038] The travel distance coefficient, the workload coefficient and the function evaluation coefficient are normalized, and the dispatch coefficient of the staff is calculated by the normalized travel distance coefficient, the workload coefficient and the function evaluation coefficient;

[0039] The staff corresponding to the maximum dispatch coefficient in the initial dispatchable staff set is determined as the final dispatchable staff, and they are sent to the target park area to monitor the environment of the target park area.

[0040] In a second aspect of the present invention, a public management space monitoring device based on a drone image vision algorithm is proposed, the device comprising:

[0041] Image quality feature module: obtain an image set of the target park area through a drone to obtain a first image set, and perform feature extraction on each image in the first image set to obtain image quality feature data of the image;

[0042] Image quality grade module: for each image in the first image set, determine the preset quality grade of the image according to the image quality feature data, and screen out images that meet the required preset quality grade as the second image set; the preset quality grades include high grade, medium grade and low grade, and the required preset quality grade is high grade;

[0043] Environmental problem type module: inputs the images in the second image set into a preset environmental problem type model, and outputs an environmental problem type set for the target park area;

[0044] Dispatch module: Determine the staff that can be dispatched based on the set of environmental problem types and staff status data to monitor the environment of the target park area.

[0045] Optionally, the image quality feature module includes:

[0046] Edge strength image module: calculates the horizontal gradient and the vertical gradient of each image in the first image set, and obtains the edge strength image according to the horizontal gradient and the vertical gradient;

[0047] Transformation module: calculating the second-order derivative of the edge intensity image to obtain a transformed image;

[0048] Edge clarity module: calculates the average pixel value of the transformed image, calculates the variance of the changed image according to the average pixel value, and uses the variance of the changed image as the edge clarity of the image.

[0049] Optionally, the image quality feature module further includes:

[0050] Weight value module: obtain several key areas of the target park area and determine the important weight value of each key area;

[0051] Weighted area module: obtains the area of ​​different key areas corresponding to each image in the first image set, and multiplies the area of ​​each key area by the corresponding weight value to obtain the weighted area of ​​the key area;

[0052] Coverage module: add up the weighted areas of all key areas of each image to get the total weighted area, and divide the total weighted area by the area of ​​the image to get the key area coverage of the image;

[0053] Light uniformity module: convert each image in the first image set into a grayscale image, calculate the average value of the grayscale values ​​of all pixels in the grayscale image as the mean, calculate the standard deviation of the image based on the mean, and take the inverse of the standard deviation as the light uniformity of the image.

[0054] Optionally, the image quality level module includes:

[0055] Input item module: The edge clarity, key area coverage and illumination uniformity of the image are defined as the input items of the fuzzy rules and are classified into different fuzzy sets respectively;

[0056] Output variable module: takes the preset quality level as the output variable and divides it into fuzzy sets;

[0057] Rule-making module: formulates fuzzy rules to describe the influence of different input variables on output variables;

[0058] Reasoning module: performs fuzzy reasoning based on fuzzy rules and outputs the preset quality level of the image.

[0059] Optionally, the environmental problem type module includes:

[0060] A preliminary determination module: determining a preliminary set of dispatchable workers according to the type of environmental problem, and obtaining status data of each worker in the preliminary set of dispatchable workers;

[0061] Driving distance module: obtain the shortest drivable distance from each staff member to the target environment area, as well as the average speed of the vehicle and the preset normal driving speed corresponding to the shortest drivable distance, and obtain the staff member's driving distance coefficient; the calculation formula is: HK = ws / yh × ut, where HK is the driving distance coefficient, ws, yh and ut are the average speed of the vehicle, the preset normal driving speed and the shortest drivable distance respectively;

[0062] Workload module: Obtain the current unfinished task volume, maximum task load volume and free time from the last task completion of each staff member, and obtain the staff member's workload coefficient. The calculation formula is: GF = ds × fg / th, where GF is the workload coefficient, th, ds and fg are the current unfinished task volume, maximum task load volume and free time from the last task completion, respectively;

[0063] Function evaluation module: obtain the number of work times and environmental problem monitoring accuracy of each staff member based on historical data, and perform weighted summation of the number of work times and environmental problem monitoring accuracy to obtain the staff member's function evaluation coefficient;

[0064] Dispatch coefficient module: normalizes the travel distance coefficient, workload coefficient and function evaluation coefficient, and calculates the dispatch coefficient of the staff member through the normalized travel distance coefficient, workload coefficient and function evaluation coefficient;

[0065] Final dispatch module: The staff corresponding to the maximum dispatch coefficient in the initial dispatchable staff set is determined as the final dispatchable staff, and they are sent to the target park area to monitor the environment of the target park area.

[0066] Beneficial effects of the present invention:

[0067] The present invention proposes a public management space monitoring method and device based on a drone image vision algorithm. The drone is used to shoot and analyze the target park area to obtain possible environmental problems in the target park area, and then appropriate staff are dispatched to further monitor the environment of the target park area based on the possible environmental problems in the target park area. In this way, the staff can be managed and dispatched in a timely and reasonable manner according to the environmental status in the park, so that the monitoring effect of the park environment by the staff in the park is higher, the staff can monitor the environmental problems in the park in a timely and accurate manner, and reduce the impact on the overall environmental quality of the park and the visitor experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] The present invention will be further described below in conjunction with the accompanying drawings.

[0069] Figure 1 A flowchart of a public management space monitoring method based on a drone image vision algorithm provided by the present invention;

[0070] Figure 2 A framework diagram of the public management space monitoring device based on the drone image vision algorithm provided by the present invention. DETAILED DESCRIPTION

[0071] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0072] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without making any creative work shall fall within the scope of protection of the present invention.

[0073] The embodiment of the present invention provides a method for monitoring public management space based on drone image vision algorithm. Figure 1 , Figure 1 A flowchart of a public management space monitoring method based on a drone image vision algorithm provided in an embodiment of the present invention. The method comprises the following steps:

[0074] Acquire an image set of a target park area through a drone to obtain a first image set, and perform feature extraction on each image in the first image set to obtain image quality feature data of the image;

[0075] For each image in the first image set, determine a preset quality level of the image according to the image quality characteristic data, and select images meeting the required preset quality level as the second image set;

[0076] Inputting the images in the second image set into a preset environmental problem type model, and outputting an environmental problem type set for the target park area;

[0077] The staff that can be finally dispatched are determined based on the set of environmental problem types and staff status data to monitor the environment of the target park area.

[0078] Based on the public management space monitoring method based on drone image vision algorithm provided by the embodiment of the present invention, through the above-mentioned method, it is possible to timely and reasonably manage and dispatch the staff according to the environmental status in the park, so that the monitoring effect of the park environment by the staff in the park is higher, and the staff can timely and accurately monitor the environmental problems in the park, reducing the impact on the overall environmental quality of the park and the visitor experience.

[0079] It should be noted that the target park area refers to one of several areas that have been divided in advance in the park. By shooting the target area with a drone, the environmental status of the target area can be obtained more comprehensively and quickly, and then subsequent analysis can be carried out; compared with the traditional simple reliance on manual methods to monitor the park's environment, based on the environmental images of the target area shot by drones and then analyzed, environmental problems can be discovered more efficiently and quickly. Drones can cover a wide area, obtain high-resolution images, and transmit data in real time, so that managers can understand the environmental conditions of each area in a timely manner, thereby optimizing scheduling and management measures. This technology not only improves the accuracy and efficiency of monitoring, but also reduces the burden on staff, ensuring the long-term cleanliness and beauty of the park environment.

[0080] It should be noted that the image set of the target area is obtained by the drone and the image set is preprocessed. The preprocessing operations include denoising, color enhancement, contrast adjustment, etc. The purpose is to improve the quality of the image and make the subsequent feature extraction more accurate. The specific image preprocessing process is an existing technology and will not be repeated here. The preprocessed image set is used as the first image set to facilitate subsequent analysis.

[0081] In one embodiment, feature extraction is performed on each image in the first image set to obtain image quality feature data of the image, including:

[0082] The image quality characteristic data includes the edge clarity of the image. The edge clarity is obtained by:

[0083] Calculating the gradient in the horizontal direction and the gradient in the vertical direction of each image in the first image set, and obtaining an edge intensity image according to the gradient in the horizontal direction and the gradient in the vertical direction;

[0084] Calculate the second-order derivative of the edge intensity image to obtain a transformed image;

[0085] The average pixel value of the transformed image is calculated, and the variance of the changed image is calculated based on the average pixel value, and the variance of the changed image is used as the edge clarity of the image.

[0086] It should be noted that when calculating the horizontal and vertical gradients of an image, the Sobel operator can be used to calculate the horizontal and vertical gradients of an image. The calculation formula is:

[0087]

[0088] In the formula, G x is the horizontal gradient of the image, G y is the vertical gradient of the image, and I is the original input image;

[0089] The edge strength image is composed of horizontal and vertical gradients, and the calculation formula can be:

[0090]

[0091] Where G is the edge strength image; the edge strength image shows the edge strength at each position in the image.

[0092] Calculate the second-order derivative of the edge intensity image to obtain the transformed image. The Laplacian operator can be used to calculate the edge intensity image to obtain the transformed image L. The calculation formula is:

[0093]

[0094] In the formula, Represents the second-order derivative of the edge intensity image in the horizontal direction; Represents the second-order derivative of the edge intensity image in the vertical direction; the transformed image reflects the second-order changes in the edge intensity image and reveals the details and changes of the edge;

[0095] Calculate the average pixel value of the transformed image. The calculation formula can be:

[0096]

[0097] Where μL is the average pixel value of the transformed image, M is the height of the image (number of rows), N is the width of the image (number of columns), and L is ij is the pixel value of the transformed image at position (i, j);

[0098] Calculate the variance of the change image. The calculation formula can be:

[0099]

[0100] σ 2 L is the variance of the change image; the variance of the change image is used as the edge clarity of the image.

[0101] In one implementation, by first calculating the edge intensity image and then applying the Laplacian operator thereto, the edge features in the image may be better captured, and the image may be more robust to noise, and the edge clarity of the calculated image is more consistent with the edge clarity in the actual image.

[0102] It should be noted that the greater the edge clarity, the higher the image quality level, because higher edge clarity usually indicates that there are more details and changes in the transformed image, which means that there may be more significant edges and textures in the original image. In images with prominent edges and details, high variance reflects higher clarity, and the corresponding image can more accurately reflect the details in the target park area environment, making the analysis of the target park area environment type more accurate and detailed, and more in line with the actual environment of the target park area.

[0103] In one embodiment, the image quality feature data further includes key area coverage and illumination uniformity;

[0104] Methods for obtaining key area coverage include:

[0105] Obtain several key areas of the target park area and determine the important weight value of each key area;

[0106] Obtaining the area of ​​different key regions corresponding to each image in the first image set, and multiplying the area of ​​each key region by the corresponding weight value to obtain the weighted area of ​​the key region;

[0107] The weighted areas of all key areas of each image are added together to obtain the total weighted area, and the total weighted area is divided by the area of ​​the image to obtain the key area coverage of the image;

[0108] Methods for obtaining light uniformity include:

[0109] Each image in the first image set is converted into a grayscale image, and the average value of the grayscale values ​​of all pixels in the grayscale image is calculated as the mean, and the standard deviation of the image is calculated according to the mean, and the reciprocal of the standard deviation is used as the illumination uniformity of the image.

[0110] It should be noted that the different key areas of the target park area are divided in advance by professional staff based on actual conditions, and the corresponding importance weight values ​​of different key areas are also set by professionals, which are set specifically according to the actual environment of the park and are not limited.

[0111] It should be noted that the greater the coverage of key areas, the higher the quality level of the image, because a higher coverage of key areas usually means that the image can cover and display more important areas in the target park area. The environment corresponding to these areas is often the focus of park management and maintenance. Through high-coverage images, the state of the park environment can be monitored and evaluated more comprehensively and accurately, which helps to discover and solve problems in a timely manner. In addition, a higher coverage of key areas also means that the image provides more information, thereby improving the efficiency and accuracy of environmental monitoring.

[0112] In one implementation method, through the above method, the key area coverage rate obtained is more accurate, more in line with the actual situation of the actual target park area, and can more comprehensively analyze the actual environmental status of the target park area.

[0113] It should be noted that the calculation steps of the illumination uniformity of the image are:

[0114] First, the image is converted into a grayscale image, and then the average grayscale value of all pixels in the grayscale image is calculated as the mean, and the standard deviation of the image is calculated based on the mean. The calculation formula can be:

[0115]

[0116] Where σ is the standard deviation, M is the height of the image (number of rows), N is the width of the image (number of columns), and I gray (x,y) represents the gray value of pixel (x,y), μ is the mean, and the inverse of the standard deviation is taken as the illumination uniformity of the image.

[0117] It should be noted that the greater the illumination uniformity of an image, the higher the image quality level, because a higher illumination uniformity of an image usually means that the image can present a more uniform illumination effect, avoiding areas that are too bright or too dark, thereby ensuring the overall visual effect and detail performance of the image. This is especially important for environmental monitoring, because images with uniform illumination can more accurately reflect the actual environmental status and reduce misjudgments and omissions caused by uneven illumination.

[0118] In one implementation method, through the above method, the obtained illumination uniformity is more accurate, more in line with the actual situation of the target park area, and can more comprehensively analyze the actual environmental status of the target park area.

[0119] In one embodiment, determining a preset quality level of an image according to the image quality characteristic data includes:

[0120] Image quality feature data includes image edge clarity, key area coverage, and illumination uniformity;

[0121] The edge definition, key area coverage and illumination uniformity of the image are defined as the input items of the fuzzy rules and classified into different fuzzy sets respectively;

[0122] The preset quality level is used as the output variable and divided into fuzzy sets;

[0123] Formulate fuzzy rules to describe the influence of different input variables on output variables;

[0124] Perform fuzzy reasoning based on fuzzy rules and output the preset quality level of the image.

[0125] It should be noted that the preset quality levels include high, medium and low;

[0126] The preset quality level of the image output in the above manner can be, for example:

[0127] Define the structure of fuzzy rules:

[0128] Each fuzzy rule consists of an "if-then" structure.

[0129] The "if" part contains the fuzzy sets of input variables.

[0130] The "then" part describes the fuzzy set of output variables (preset quality levels).

[0131] Formulate fuzzy rules, such as:

[0132] Rule 1: If edge clarity is "High", key area coverage is "High Coverage", and illumination uniformity is "Very Uniform", then the preset quality level is "High".

[0133] Rule 2: If edge clarity is "Medium", key area coverage is "Medium Coverage", and illumination uniformity is "Uniform", then the preset quality level is "Medium".

[0134] Rule 3: If edge clarity is "Low" or key area coverage is "Low Coverage" or illumination uniformity is "Uneven", the preset quality level is "Low Level".

[0135] Perform fuzzy reasoning;

[0136] Fuzzification: Map the actual values ​​of edge sharpness, key area coverage, and illumination uniformity of the input image to the corresponding fuzzy sets. For example, if the edge sharpness is 0.7, it can be mapped to the "medium" set, while if the key area coverage is 0.3, it can be mapped to the "low coverage" set.

[0137] Apply fuzzy rules: Based on the fuzzy values ​​of the input variables, use fuzzy rules to generate the fuzzy values ​​of the rule outputs. For example, if edge clarity is "low", key area coverage is "low coverage", and illumination uniformity is "uneven", then according to rule 3, the fuzzy value of the preset quality level "low level" will increase.

[0138] Synthesize fuzzy results: The outputs of all rules are combined into a fuzzy output set. For each preset quality level (high, medium, low), a corresponding fuzziness value will be obtained, which reflects the possibility of that level.

[0139] Defuzzification:

[0140] Use an appropriate defuzzification method (such as the centroid method) to convert the fuzzy set into a specific quality level value. The centroid method usually calculates the weighted average of the center points of all fuzzy sets to obtain a final clarity value.

[0141] Result output: According to the result after defuzzification, the final preset quality level of the image is determined. For example, if the result after defuzzification shows that the blurriness of "low level" is the largest, the preset quality level of the image is "low level".

[0142] In one implementation, by formulating fuzzy rules and fuzzy reasoning, the preset quality level of the image can be effectively determined based on the edge clarity, key area coverage and illumination uniformity of the image. This method can handle uncertainty and ambiguity and provide a reasonable image quality assessment.

[0143] The images with a high preset quality level in the first image set are defined as images with a desired preset quality level, and all the images with the desired preset quality level in the first image set are reassembled to obtain a second image set.

[0144] In one implementation method, the second image set obtained through the above method can ensure comprehensive coverage and uniform lighting of key areas, thereby improving the accuracy and efficiency of environmental monitoring, timely discovering and solving environmental problems in the park, optimizing resource scheduling, and improving the overall level of park management.

[0145] In one embodiment, the images in the second image set are input into a preset environmental problem type model, and the environmental problem type set of the target park area is outputted, including:

[0146] Use a preset environmental problem type model to analyze each image in the second image set. The preset environmental problem type model can be a trained convolutional neural network (CNN) or other image classification model. The model has been trained with a large amount of labeled data and can identify environmental problems in the park. The preset environmental problem type model outputs the types of environmental problems that may exist in the target park area analyzed for each image and their probabilities. For example, the environmental types and problems of an image analyzed by the preset environmental problem type model are: garbage accumulation: 30%, excessive vegetation growth: 35%, and water pollution 35%. Garbage accumulation, excessive vegetation growth, and water pollution are used as the environmental problem type set for the image. The environmental problem types corresponding to all images in the second image set are combined to determine the environmental problem type set for the target park area.

[0147] In one implementation method, through the above method, the types of environmental problems that may exist in the target park area can be obtained based on the images taken by the drone, and the types of environmental problems that may exist in the target park area can be relatively comprehensive and accurate, and the subsequent dispatch and management of staff can be determined based on the set of environmental problem types.

[0148] In one embodiment, determining the workers that can be dispatched eventually based on the set of environmental problem types and worker status data includes:

[0149] Determine a preliminary set of dispatchable workers according to the type of environmental problem, and obtain status data of each worker in the preliminary set of dispatchable workers;

[0150] The status data of the staff include the travel distance coefficient, workload coefficient and function evaluation coefficient;

[0151] The method for obtaining the driving distance coefficient is:

[0152] The shortest drivable distance from each staff member to the target environmental area, as well as the average speed of the vehicle and the preset normal driving speed corresponding to the shortest drivable distance are obtained to obtain the staff member's driving distance coefficient; the calculation formula is: HK = ws / yh × ut, where HK is the driving distance coefficient, ws, yh and ut are the average speed of the vehicle, the preset normal driving speed and the shortest drivable distance respectively;

[0153] The workload factor is obtained as follows;

[0154] Obtain the current unfinished task amount, maximum task load, and free time from the last task completion of each staff member, and obtain the staff member's workload coefficient. The calculation formula is: GF = ds × fg / th, where GF is the workload coefficient, th, ds, and fg are the current unfinished task amount, maximum task load, and free time from the last task completion, respectively;

[0155] The method for obtaining the function evaluation coefficient is as follows:

[0156] The number of times each staff member worked and the accuracy of monitoring environmental issues were obtained based on historical data, and the number of times each staff member worked and the accuracy of monitoring environmental issues were weighted and summed to obtain the staff member's function evaluation coefficient;

[0157] The travel distance coefficient, the workload coefficient and the function evaluation coefficient are normalized, and the dispatch coefficient of the staff is calculated by the normalized travel distance coefficient, the workload coefficient and the function evaluation coefficient;

[0158] The staff corresponding to the maximum dispatch coefficient in the initial dispatchable staff set is determined as the final dispatchable staff, and they are sent to the target park area to monitor the environment of the target park area.

[0159] It should be noted that the shortest drivable distance from each staff member to the target environmental area can be obtained in real time through map services (such as Baidu Maps, Amap), and the average speed of the vehicle corresponding to the shortest drivable distance can be obtained through the corresponding real-time traffic management system, or other methods, without specific restrictions; in addition, the preset normal driving speed is set by professionals within the budget according to the actual situation, and there is no specific restriction. The staff's current unfinished task volume, maximum task load, free time from the last task completion, staff work times and environmental problem monitoring accuracy can be obtained based on the staff's work log, or other methods, without specific restrictions.

[0160] It should be noted that the function evaluation coefficient of the staff member can be calculated by the following formula by weighted summing the number of work and the accuracy of environmental problem monitoring, for example: TG = a1 × lm + a2 × pg, where TG is the function evaluation coefficient, lm and pg are the number of work and the accuracy of environmental problem monitoring, respectively, and a1 and a2 are the proportional coefficients of the number of work and the accuracy of environmental problem monitoring, and a1 and a2 are both greater than 0;

[0161] It should be noted that, for example, the present invention can use the following formula to calculate the dispatch coefficient of the staff, and the calculation expression is: Dxs=f1×HK+f2×GF+f3×TG, wherein Dxs is the dispatch coefficient, HK, GF, TG are respectively the driving distance coefficient, the workload coefficient and the function evaluation coefficient, f1, f2, f3 are respectively the proportional coefficients of the driving distance coefficient, the workload coefficient and the function evaluation coefficient, and f1, f2, f3 are all greater than 0.

[0162] It should be noted that the driving distance coefficient is used to measure the distance and driving efficiency of employees to reach the target area, ensuring that the selected employees can arrive at the site in a shorter time, which helps to improve response speed; the workload coefficient reflects the workload and fatigue level of employees, avoiding sending overly tired employees to tasks, which helps to maintain employee work efficiency and accuracy and reduce errors caused by fatigue; the functional evaluation coefficient is based on the employee's historical work performance and problem monitoring accuracy, ensuring that employees with higher professional capabilities are selected, thereby improving the quality and reliability of environmental problem monitoring.

[0163] It should be noted that the initial set of dispatchable staff members is determined based on the type of environmental problem. The staff members who can solve all these problems are first screened out according to the type of environmental problem. Specifically, the skills and experience of the staff members can be matched to ensure that they have the ability to deal with specific environmental problems. Specifically, the historical work records and areas of expertise of the staff members can be analyzed to ensure that the selected personnel can effectively deal with all environmental problems in the target park area, thereby improving the efficiency and effectiveness of problem solving. Other methods are also possible and are not limited to specific methods.

[0164] In one implementation, through the above method, according to the calculation of the dispatch coefficient of the employee, the most suitable employee can be scientifically and reasonably selected for task dispatch, and the environment of the target park area can be further accurately and meticulously monitored. In this way, the staff can be managed and dispatched in a timely and reasonable manner according to the environmental status in the park, so that the staff has a higher monitoring effect on the park environment.

[0165] Based on the same inventive concept, the embodiment of the present invention also provides a public management space monitoring device based on drone image vision algorithm. Figure 2 , Figure 2 A schematic diagram of a framework of a public management space monitoring device based on a drone image vision algorithm provided by an embodiment of the present invention, the device comprising:

[0166] Image quality feature module: obtain an image set of the target park area through a drone to obtain a first image set, and perform feature extraction on each image in the first image set to obtain image quality feature data of the image;

[0167] Image quality grade module: for each image in the first image set, determining a preset quality grade of the image according to the image quality characteristic data, and screening out images meeting the required preset quality grade as the second image set;

[0168] Environmental problem type module: inputs the images in the second image set into a preset environmental problem type model, and outputs an environmental problem type set for the target park area;

[0169] Dispatch module: Determine the staff that can be dispatched based on the set of environmental problem types and staff status data to monitor the environment of the target park area.

[0170] Based on the public management space monitoring device based on the drone image vision algorithm provided by the embodiment of the present invention, through the above-mentioned method, it is possible to timely and reasonably manage and dispatch the staff according to the environmental status in the park, so that the monitoring effect of the park environment by the staff in the park is higher, and the staff can timely and accurately monitor the environmental problems in the park, reducing the impact on the overall environmental quality of the park and the visitor experience.

[0171] In one embodiment, the image quality feature module includes:

[0172] Edge strength image module: calculates the horizontal gradient and the vertical gradient of each image in the first image set, and obtains the edge strength image according to the horizontal gradient and the vertical gradient;

[0173] Transformation module: calculate the second-order derivative of the edge intensity image to obtain the transformed image;

[0174] Edge clarity module: Calculate the average pixel value of the transformed image, calculate the variance of the changed image based on the average pixel value, and use the variance of the changed image as the edge clarity of the image.

[0175] In one embodiment, the image quality feature module further includes:

[0176] Weight value module: obtain several key areas of the target park area and determine the important weight value of each key area;

[0177] Weighted area module: obtains the area of ​​different key areas corresponding to each image in the first image set, and multiplies the area of ​​each key area by the corresponding weight value to obtain the weighted area of ​​the key area;

[0178] Coverage module: add up the weighted areas of all key areas of each image to get the total weighted area, and divide the total weighted area by the area of ​​the image to get the key area coverage of the image;

[0179] Light uniformity module: convert each image in the first image set into a grayscale image, calculate the average value of the grayscale values ​​of all pixels in the grayscale image as the mean, calculate the standard deviation of the image based on the mean, and take the inverse of the standard deviation as the light uniformity of the image.

[0180] In one embodiment, the image quality rating module includes:

[0181] Input item module: The edge clarity, key area coverage and illumination uniformity of the image are defined as the input items of the fuzzy rules and are classified into different fuzzy sets respectively;

[0182] Output variable module: takes the preset quality level as the output variable and divides it into fuzzy sets;

[0183] Rule-making module: formulates fuzzy rules to describe the influence of different input variables on output variables;

[0184] Reasoning module: performs fuzzy reasoning based on fuzzy rules and outputs the preset quality level of the image.

[0185] In one embodiment, the environmental issue type module includes:

[0186] Preliminary determination module: determines the preliminary set of dispatchable workers according to the type of environmental problem, and obtains the status data of each worker in the preliminary set of dispatchable workers;

[0187] Driving distance module: obtain the shortest drivable distance from each staff member to the target environment area, as well as the average speed of the vehicle and the preset normal driving speed corresponding to the shortest drivable distance, and obtain the staff member's driving distance coefficient; the calculation formula is: HK = ws / yh × ut, where HK is the driving distance coefficient, ws, yh and ut are the average speed of the vehicle, the preset normal driving speed and the shortest drivable distance respectively;

[0188] Workload module: Obtain the current unfinished task volume, maximum task load volume and free time from the last task completion of each staff member, and obtain the staff member's workload coefficient. The calculation formula is: GF = ds × fg / th, where GF is the workload coefficient, th, ds and fg are the current unfinished task volume, maximum task load volume and free time from the last task completion, respectively;

[0189] Function evaluation module: obtain the number of work times and environmental problem monitoring accuracy of each staff member based on historical data, and perform weighted summation of the number of work times and environmental problem monitoring accuracy to obtain the staff member's function evaluation coefficient;

[0190] Dispatch coefficient module: normalizes the travel distance coefficient, workload coefficient and function evaluation coefficient, and calculates the dispatch coefficient of the staff member through the normalized travel distance coefficient, workload coefficient and function evaluation coefficient;

[0191] Final dispatch module: The staff corresponding to the maximum dispatch coefficient in the initial dispatchable staff set is determined as the final dispatchable staff, and they are sent to the target park area to monitor the environment of the target park area.

[0192] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A public management space monitoring method based on drone image vision algorithm, characterized in that: The following steps are involved: Acquire an image set of a target park area through a drone to obtain a first image set, and perform feature extraction on each image in the first image set to obtain image quality feature data of the image; For each image in the first image set, determine a preset quality level of the image according to the image quality feature data, and select images that meet the required preset quality level as the second image set; the preset quality levels include high level, medium level and low level, and the required preset quality level is high level; Inputting the images in the second image set into a preset environmental problem type model, and outputting an environmental problem type set for the target park area; The staff that can be finally dispatched are determined based on the set of environmental problem types and staff status data to monitor the environment of the target park area.

2. The public management space monitoring method based on drone image vision algorithm according to claim 1 is characterized in that: Feature extraction is performed on each image in the first image set to obtain image quality feature data of the image including: The image quality characteristic data includes the edge clarity of the image, and the edge clarity is obtained by: Calculating the gradient in the horizontal direction and the gradient in the vertical direction of each image in the first image set, and obtaining an edge intensity image according to the gradient in the horizontal direction and the gradient in the vertical direction; Calculating the second-order derivative of the edge intensity image to obtain a transformed image; The average pixel value of the transformed image is calculated, and the variance of the changed image is calculated according to the average pixel value, and the variance of the changed image is used as the edge clarity of the image.

3. The public management space monitoring method based on drone image vision algorithm according to claim 2 is characterized in that: The image quality feature data also includes key area coverage and illumination uniformity; The method for obtaining the key area coverage includes: Obtain several key areas of the target park area and determine the important weight value of each key area; Obtaining the area of ​​different key regions corresponding to each image in the first image set, and multiplying the area of ​​each key region by the corresponding weight value to obtain the weighted area of ​​the key region; The weighted areas of all key areas of each image are added together to obtain the total weighted area, and the total weighted area is divided by the area of ​​the image to obtain the key area coverage of the image; The method for obtaining the illumination uniformity includes: Each image in the first image set is converted into a grayscale image, and the average value of the grayscale values ​​of all pixels in the grayscale image is calculated as the mean, and the standard deviation of the image is calculated according to the mean, and the reciprocal of the standard deviation is taken as the illumination uniformity of the image.

4. The public management space monitoring method based on drone image vision algorithm according to claim 1 is characterized in that: The preset quality levels of the images determined based on the image quality characteristic data include: The image quality feature data includes edge clarity, key area coverage and illumination uniformity of the image; The edge definition, key area coverage and illumination uniformity of the image are defined as the input items of the fuzzy rules and classified into different fuzzy sets respectively; The preset quality level is used as the output variable and divided into fuzzy sets; Formulate fuzzy rules to describe the influence of different input variables on output variables; Perform fuzzy reasoning based on fuzzy rules and output the preset quality level of the image.

5. The public management space monitoring method based on drone image vision algorithm according to claim 1 is characterized in that: The workers that can be dispatched based on the environmental problem type set and worker status data include: Determine a preliminary set of dispatchable workers according to the type of environmental problem, and obtain status data of each worker in the preliminary set of dispatchable workers; The staff member's status data includes a travel distance coefficient, a workload coefficient and a function evaluation coefficient; The method for obtaining the driving distance coefficient is: The shortest drivable distance from each staff member to the target environmental area, as well as the average speed of the vehicle and the preset normal driving speed corresponding to the shortest drivable distance are obtained to obtain the staff member's driving distance coefficient; the calculation formula is: HK = ws / yh × ut, where HK is the driving distance coefficient, ws, yh and ut are the average speed of the vehicle, the preset normal driving speed and the shortest drivable distance respectively; The method for obtaining the workload coefficient is: Obtain the current unfinished task amount, maximum task load, and free time from the last task completion of each staff member, and obtain the staff member's workload coefficient. The calculation formula is: GF = ds × fg / th, where GF is the workload coefficient, th, ds, and fg are the current unfinished task amount, maximum task load, and free time from the last task completion, respectively; The method for obtaining the function evaluation coefficient is as follows: The number of times each staff member worked and the accuracy of monitoring environmental issues were obtained based on historical data, and the number of times each staff member worked and the accuracy of monitoring environmental issues were weighted and summed to obtain the staff member's function evaluation coefficient; The travel distance coefficient, the workload coefficient and the function evaluation coefficient are normalized, and the dispatch coefficient of the staff is calculated by the normalized travel distance coefficient, the workload coefficient and the function evaluation coefficient; The staff corresponding to the maximum dispatch coefficient in the initial dispatchable staff set is determined as the final dispatchable staff, and they are sent to the target park area to monitor the environment of the target park area.

6. A public management space monitoring device based on drone image vision algorithm, characterized in that: The device comprises: Image quality feature module: obtain an image set of the target park area through a drone to obtain a first image set, and perform feature extraction on each image in the first image set to obtain image quality feature data of the image; Image quality grade module: for each image in the first image set, determine the preset quality grade of the image according to the image quality feature data, and screen out images that meet the required preset quality grade as the second image set; the preset quality grades include high grade, medium grade and low grade, and the required preset quality grade is high grade; Environmental problem type module: inputs the images in the second image set into a preset environmental problem type model, and outputs an environmental problem type set for the target park area; Dispatch module: Determine the staff that can be dispatched based on the set of environmental problem types and staff status data to monitor the environment of the target park area.

7. The public management space monitoring device based on drone image vision algorithm according to claim 6 is characterized in that: The image quality feature module comprises: Edge strength image module: calculates the horizontal gradient and the vertical gradient of each image in the first image set, and obtains the edge strength image according to the horizontal gradient and the vertical gradient; Transformation module: calculating the second-order derivative of the edge intensity image to obtain a transformed image; Edge clarity module: calculates the average pixel value of the transformed image, calculates the variance of the changed image according to the average pixel value, and uses the variance of the changed image as the edge clarity of the image.

8. The public management space monitoring device based on drone image vision algorithm according to claim 7 is characterized in that: The image quality feature module also includes: Weight value module: obtain several key areas of the target park area and determine the important weight value of each key area; Weighted area module: obtains the area of ​​different key areas corresponding to each image in the first image set, and multiplies the area of ​​each key area by the corresponding weight value to obtain the weighted area of ​​the key area; Coverage module: add up the weighted areas of all key areas of each image to get the total weighted area, and divide the total weighted area by the area of ​​the image to get the key area coverage of the image; Light uniformity module: convert each image in the first image set into a grayscale image, calculate the average value of the grayscale values ​​of all pixels in the grayscale image as the mean, calculate the standard deviation of the image based on the mean, and take the inverse of the standard deviation as the light uniformity of the image.

9. The public management space monitoring device based on drone image vision algorithm according to claim 6 is characterized in that: The image quality grade module includes: Input item module: The edge clarity, key area coverage and illumination uniformity of the image are defined as the input items of the fuzzy rules and are classified into different fuzzy sets respectively; Output variable module: takes the preset quality level as the output variable and divides it into fuzzy sets; Rule-making module: formulates fuzzy rules to describe the influence of different input variables on output variables; Reasoning module: performs fuzzy reasoning based on fuzzy rules and outputs the preset quality level of the image.

10. The public management space monitoring device based on drone image vision algorithm according to claim 6 is characterized in that: The environmental problem type module includes: A preliminary determination module: determining a preliminary set of dispatchable workers according to the type of environmental problem, and obtaining status data of each worker in the preliminary set of dispatchable workers; Driving distance module: obtain the shortest drivable distance from each staff member to the target environment area, as well as the average speed of the vehicle and the preset normal driving speed corresponding to the shortest drivable distance, and obtain the staff member's driving distance coefficient; the calculation formula is: HK = ws / yh × ut, where HK is the driving distance coefficient, ws, yh and ut are the average speed of the vehicle, the preset normal driving speed and the shortest drivable distance respectively; Workload module: Obtain the current unfinished task volume, maximum task load volume and free time from the last task completion of each staff member, and obtain the staff member's workload coefficient. The calculation formula is: GF = ds × fg / th, where GF is the workload coefficient, th, ds and fg are the current unfinished task volume, maximum task load volume and free time from the last task completion, respectively; Function evaluation module: obtain the number of work times and environmental problem monitoring accuracy of each staff member based on historical data, and perform weighted summation of the number of work times and environmental problem monitoring accuracy to obtain the staff member's function evaluation coefficient; Dispatch coefficient module: normalizes the travel distance coefficient, workload coefficient and function evaluation coefficient, and calculates the dispatch coefficient of the staff member through the normalized travel distance coefficient, workload coefficient and function evaluation coefficient; Final dispatch module: The staff corresponding to the maximum dispatch coefficient in the initial dispatchable staff set is determined as the final dispatchable staff, and they are sent to the target park area to monitor the environment of the target park area.

Images