Automatic Video Inspection System Based on GIS

Abstract

The present invention discloses an automatic video inspection system based on GIS, which relates to the technical field of intelligent monitoring, and solves the technical problems that it is difficult to import video data, geographic spatial data and environmental parameters into a GIS platform for spatial overlay analysis; secondly, it is difficult to identify pest-infested crops and analyze the variability of pest-infested crops; then, based on environmental parameters and video data, it is difficult to analyze the growth index and soil pollution rate of crops; finally, based on various data of farmland, it is difficult to plan inspections of farmland. The present invention collects farmland data through a data acquisition module, and pre-processes the collected data through a data processing module. The farmland is divided into multiple planting areas through a farmland division module. The geospatial data, environmental data and video data of the farmland are analyzed through a data analysis module; the rating of the farmland is analyzed through a farmland rating inspection module, and the farmland is planned for inspection.

Description

Technical Field

[0001] The invention belongs to the technical field of intelligent monitoring, and in particular is an automatic video inspection system based on GIS. Background Art

[0002] With the acceleration of urbanization and the continuous expansion of infrastructure, traditional manual inspection methods can no longer meet the needs of efficient, real-time and comprehensive monitoring. Manual inspections are not only time-consuming and labor-intensive, but also difficult to detect potential problems in a timely manner.

[0003] The existing GIS-based automatic video inspection system has the following problems: first, it is difficult to import video data, geospatial data and environmental parameters into the GIS platform for spatial overlay analysis; second, it is difficult to identify pest-infested crops and analyze the variability of pest-infested crops; then, it is difficult to analyze the growth index of crops and the soil pollution rate based on environmental parameters and video data; finally, it is difficult to plan inspections of farmland based on various data of the farmland. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes an automatic video inspection system based on GIS, which is used to solve the following technical problems:

[0005] First, it is difficult to import video data, geospatial data and environmental parameters into the GIS platform for spatial overlay analysis. Second, it is difficult to identify pest-infested crops and analyze their variability. Next, based on environmental parameters and video data, it is difficult to analyze crop growth indexes and soil contamination rates. Finally, based on various farmland data, it is difficult to plan inspections of farmland.

[0006] To solve the above problems, the first aspect of the present invention provides a GIS-based automatic video inspection system, comprising the following modules:

[0007] Data collection module: collects geospatial data, environmental data and video data of farmland;

[0008] Data processing module: pre-processing the collected video data, including: denoising and image enhancement; pre-processing the collected geospatial data and environmental parameters, including: cleaning and normalization;

[0009] Farmland division module: According to different crop growth requirements, the farmland is divided into n types of crop planting areas, and the i-th planting area is marked as the x-th i types of crops planted, where i∈{1,2,...,n};

[0010] Data analysis module: import the pre-processed video data, geospatial data and environmental parameters into the GIS platform for spatial overlay analysis; use deep learning algorithms to identify pests and diseases on the processed video data, analyze and identify the xth i Calculate the growth rate and coverage rate of pest-infested crops according to the analysis results and calculate the pest-infested crop variation rate through comprehensive analysis; Analyze and calculate the growth index of crops according to the collected and processed environmental parameters; Analyze the xth video data of the farmland according to the collected and processed video data of the farmland i the soil contamination rate in the areas where the crops are grown;

[0011] Farmland rating and inspection module: Based on the pest and crop variability, growth index and soil pollution rate of the farmland, the farmland is comprehensively rated and the number of farmland inspections is dynamically adjusted based on the results.

[0012] As a further solution of the present invention: collecting geospatial data, environmental data and video data of farmland includes the following steps:

[0013] The geospatial data collection uses a global positioning system (GPS) device to obtain geographical information of the farmland, including: longitude and latitude coordinates, farmland boundaries, farmland area and farmland shape;

[0014] Environmental sensors are deployed in farmland, including soil temperature and humidity sensors, air temperature and humidity sensors, light sensors, and carbon dioxide concentration sensors, to monitor the environmental parameters of farmland in real time, including soil temperature and humidity, air temperature and humidity, light intensity, and carbon dioxide concentration;

[0015] Use drones equipped with high-definition cameras, plan the flight route and altitude of the drones through GIS, and plan the frequency of farmland inspections.

[0016] As a further solution of the present invention: using a drone equipped with a high-definition camera, planning the flight route and altitude of the drone by using GIS, and planning the frequency of farmland inspection, including the following steps:

[0017] According to the performance parameters of the drone, including endurance, flight speed, and the type of sensors carried, set the flight altitude, flight speed, and shooting angle of the drone in the GIS software;

[0018] Use GIS to generate the optimal flight route covering the entire farmland area based on the shape and size of the farmland; set multiple take-off and landing points based on the drone's flight time and shooting requirements; perform simulated flight in GIS to check whether the flight route is reasonable and whether there are any safety hazards or areas that cannot be covered;

[0019] The inspection frequency of farmland is planned according to the month, including: once a week from March to May; twice a week from June to August; twice a week from September to November; once a month from December to February.

[0020] As a further solution of the present invention, the pre-processed video data, geographic spatial data and environmental parameters are imported into the GIS platform for spatial overlay analysis, including the following steps:

[0021] S1: converting the preprocessed video data, geospatial data and environmental parameters into formats, including: converting video data into vector or raster data through feature extraction, converting geospatial data into vector layers with coordinate information, and converting environmental parameters into shape or surface data;

[0022] S2: Based on the geospatial data, the vector layer of the farmland’s geographic information is imported into the GIS; the data of the monitoring environmental parameters of soil temperature and humidity, air temperature and humidity, light intensity, and carbon dioxide concentration are added as attribute information to the corresponding geospatial data layer, and the video data of the farmland is imported into the GIS;

[0023] S3: Set the data to a unified coordinate system and projection for spatial overlay analysis;

[0024] S4: Analyze the monitored environmental parameters and farmland boundaries, determine the farmland area to which the monitoring point belongs, and perform point-to-polygon overlay, including: importing point data and polygon data into GIS, wherein the point data and polygon data represent the environmental parameters and farmland boundaries detected by the monitoring point, respectively; using the spatial connection or point-in-polygon function of GIS software, overlay and analyze the point data and polygon data, and generate a farmland area layer with environmental parameter information;

[0025] S5: Statistical analysis is performed on the farmland area layer with environmental parameter information superimposed on the point data and polygon data, and the field statistics function of GIS is used to calculate the x-th i The mean and variance of each environmental parameter value in the planting area of ​​​​the crops; according to the analysis results, different thresholds for different environmental parameters are set, and the selection function of the GIS software is used to screen out the farmland areas where the environmental parameters do not meet the standards according to the set thresholds, and mark the screened farmland areas with substandard environmental parameters on the map; increase the number of inspections according to the farmland areas where the environmental parameters do not meet the standards, and increase the inspection frequency planned by month from one inspection per week or month.

[0026] As a further solution of the present invention: using a deep learning algorithm to identify pests and diseases on the processed video data, analyzing and identifying the xth i Planting crops with pests and diseases, including the following steps:

[0027] Based on the pre-processed video data, the collected images are labeled to identify normal crops and crops with pests and diseases;

[0028] Use the convolutional neural network deep learning model to identify pests and diseases, and input the labeled pest and disease image data into the model for training;

[0029] The xth batch of farmland is inspected based on the video data collected in real time by the GIS-based automatic video inspection system. i Analyze crops and identify normal crops and crops with pests and diseases.

[0030] As a further solution of the present invention, the growth rate and coverage rate of the pest-infested crops are calculated according to the analysis results, and the pest-infested crops are calculated by comprehensive analysis, including the following steps:

[0031] According to the image recognition results, a patrol is carried out at time t1, and the xth i The number of pixels of pest-infested crops and normal crops of the crops; at the time t2 of the next inspection, the xth pixel identified is counted i The number of pixels of crops with pests and diseases and normal crops;

[0032] By analyzing the formula:

[0033] Get the pixel area M; where q represents the total number of pixels in the image, s x represents the area of ​​the xth pixel, x∈{1,2,...,q}; by substituting into the formula, we can get the xth pixel at time t1 and time t2 respectively. i The pixel area of ​​pest-infested crops and normal crops;

[0034] By analyzing the formula:

[0035] Get the growth rate of pest-infested crops H B (t); where Indicates that at the xth i The pixel area of ​​the pest-infested crops at the t1th moment of planting crops, Indicates that at the xth i The pixel area of ​​the pest-infested crops at time t2 of planting crops;

[0036] By analyzing the formula:

[0037]

[0038]

[0039] Get the coverage rate of pests and diseases crops F B ;in, Indicates that at the xth i The pixel area of ​​normal crops at the t1th moment of planting crops, Indicates that at the xth i The pixel area of ​​normal crops at the t2th moment of planting crops, represents the total pixel area of ​​crops planted at time t1, represents the total pixel area of ​​crops planted at time t2;

[0040] By analyzing the formula:

[0041] Get the pest and disease crop variation rate B, where H B (t) represents the growth rate of crops affected by pests and diseases, F B It indicates the coverage rate of crops with pests and diseases.

[0042] As a further solution of the present invention: analyzing and calculating the growth index of crops according to the collected and processed environmental parameters, comprising the following steps:

[0043] According to the collected and processed environmental parameters, the environmental parameter matrix (Y ab ) n×m ;

[0044] By analyzing the formula:

[0045] Get the standard environmental parameter value Y ab ′ , where a represents the xth i Planting crops, Y b represents the environmental parameter value of the bth type, Y ab Indicates the xth i The environmental parameter value of the bth crop;

[0046] By calculating the contribution of environmental parameters:

[0047] By analyzing the formula:

[0048] Get the influence coefficient K ab ′ , where G ab represents the contribution value of environmental parameters, Y ab ′ Indicates standard environmental parameter values;

[0049] By analyzing the formula:

[0050] Get the growth index of crops δ, where K ab ′ represents the influence coefficient, Y ab ′ Indicates standard environmental parameter values.

[0051] As a further solution of the present invention: according to the collected and processed video data of the farmland, the xth i The soil contamination rate of the crop planting area includes the following steps:

[0052] Based on the collected and processed video data of the farmland, an image of the soil area is extracted, the RGB value of the soil color is extracted from the image, and the average value of the RGB component of each pixel in the image is calculated;

[0053] By analyzing the formula:

[0054] The soil pollution rate R is obtained, where Red avg 、Green avg 、Blue avg are the average values ​​of red, green and blue respectively, c represents the coefficient of change of soil contamination rate, and λ represents the expected value of soil contamination rate.

[0055] As a further solution of the present invention: according to the pest and crop variation rate, growth index and soil pollution rate of the farmland, the farmland is comprehensively rated and the farmland inspection is carried out according to the results, including the following steps:

[0056] By analyzing the formula: β=B*ε1+δ*ε2+R*ε3

[0057] The comprehensive rating value of farmland β is obtained; among them, B represents the pest-infested crop variability, δ represents the growth index, and R represents the soil pollution rate; ε1, ε2, and ε3 represent the weights of the pest-infested crop variability, growth index, and soil pollution rate, respectively;

[0058] According to the comprehensive rating value of farmland, the farmland grade area is divided into high-rated value area and low-rated value area. When it is a low-rated value area, the number of inspections is increased based on the inspection frequency planned according to the month, from one inspection per week or month. Otherwise, regular inspections are carried out according to the monthly plan.

[0059] Compared with the prior art, the present invention has the following beneficial effects:

[0060] The present invention can achieve precise management of farmland by collecting and analyzing farmland geospatial data, environmental data and video data. This helps to provide suitable growth conditions for each crop according to the growth requirements of different crops, thereby improving the yield and quality of crops;

[0061] The present invention uses a deep learning algorithm to identify pests and diseases on processed video data, which can timely detect the occurrence of pests and diseases, helping farmers to take corresponding prevention and control measures in the early stages of pests and diseases and reduce losses.

[0062] The present invention can obtain the growth index of crops by analyzing and calculating environmental parameters, thereby guiding farmers to rationally apply fertilizers and irrigation, optimize the use of resources, and improve agricultural production efficiency;

[0063] By analyzing the video data of farmland, the present invention can evaluate the soil pollution rate, take timely measures to improve soil quality and protect the ecological environment of farmland; comprehensively grading farmland according to the crop variability of pests and diseases, growth index and soil pollution rate can help farmers understand the overall condition of the farmland, conduct targeted inspections and maintenance, and ensure healthy and stable production of farmland. It realizes the automation of video monitoring, improves the efficiency of monitoring and inspection, reduces manual intervention, and reduces labor costs and error rates. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0065] Figure 1 It is a system module diagram of the present invention;

[0066] Figure 2 The present invention is a flow chart of the method. DETAILED DESCRIPTION

[0067] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than 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.

[0068] See also Figure 1-2 The first embodiment of the present invention provides a GIS-based automatic video inspection system, comprising the following modules:

[0069] Data collection module: collects geospatial data, environmental data and video data of farmland;

[0070] Data processing module: pre-processing the collected video data, including: denoising and image enhancement; pre-processing the collected geospatial data and environmental parameters, including: cleaning and normalization;

[0071] Farmland division module: According to different crop growth requirements, the farmland is divided into n types of crop planting areas, and the i-th planting area is marked as the x-th i types of crops planted, where i∈{1,2,...,n};

[0072] Data analysis module: import the pre-processed video data, geospatial data and environmental parameters into the GIS platform for spatial overlay analysis; use deep learning algorithms to identify pests and diseases on the processed video data, analyze and identify the xth i Calculate the growth rate and coverage rate of pest-infested crops according to the analysis results and calculate the pest-infested crop variation rate through comprehensive analysis; Analyze and calculate the growth index of crops according to the collected and processed environmental parameters; Analyze the xth video data of the farmland according to the collected and processed video data of the farmland i the soil contamination rate in the areas where the crops are grown;

[0073] Farmland rating and inspection module: Based on the pest and crop variability, growth index and soil pollution rate of the farmland, the farmland is comprehensively rated and the number of farmland inspections is dynamically adjusted based on the results.

[0074] Specifically, the data acquisition module uses GPS and other remote sensing technologies to collect geospatial data of farmland; collects environmental parameters such as temperature, humidity, and light intensity; uses drones or fixed cameras to shoot videos of farmland to record crop growth conditions and possible pests and diseases. The data processing module performs denoising on the collected videos to enhance image quality and edge details. Preprocessing of geospatial data and environmental parameters: cleaning data, removing outliers, and normalizing. Through the farmland division module, the farmland is divided into multiple planting areas according to the growth requirements of different crops, and the corresponding crop planting type is marked for each planting area. Through the data analysis module, the preprocessed video data, geospatial data, and environmental parameters are imported into the GIS platform for spatial overlay analysis to identify specific problems in different areas. Deep learning algorithms are used to identify pests and diseases on the processed video data, and the pest and disease conditions of each crop are analyzed. The growth rate and coverage rate of pest-infested crops are calculated based on the analysis results, and the pest and disease crop variability is calculated through comprehensive analysis. According to the collected and processed environmental parameters, the growth index of crops is analyzed and calculated. Based on the collected and processed farmland video data, the soil contamination rate of each planting area is analyzed. Based on the pest and disease crop change rate, growth index and soil contamination rate, the farmland is comprehensively rated, and an inspection plan is formulated to dynamically adjust the number of farmland inspections.

[0075] In one embodiment of the present invention, collecting geospatial data, environmental data and video data of farmland includes the following steps:

[0076] The geospatial data collection uses a global positioning system (GPS) device to obtain geographical information of the farmland, including: longitude and latitude coordinates, farmland boundaries, farmland area and farmland shape;

[0077] Environmental sensors are deployed in farmland, including soil temperature and humidity sensors, air temperature and humidity sensors, light sensors, and carbon dioxide concentration sensors, to monitor the environmental parameters of farmland in real time, including soil temperature and humidity, air temperature and humidity, light intensity, and carbon dioxide concentration;

[0078] Use drones equipped with high-definition cameras, plan the flight route and altitude of the drones through GIS, and plan the frequency of farmland inspections.

[0079] Specifically, the geographical information of farmland is obtained through the global positioning system GPS device, including latitude and longitude coordinates, farmland boundaries, farmland area and farmland shape. These data provide the basis for subsequent farmland division and inspection. A series of environmental sensors are deployed in the farmland, including soil temperature and humidity sensors, air temperature and humidity sensors, light sensors and carbon dioxide concentration sensors. These sensors monitor the environmental parameters of the farmland in real time. Use drones equipped with high-definition cameras, and plan the flight route and altitude of the drones through GIS to cover the entire farmland area and avoid repeated shooting.

[0080] In one embodiment of the present invention, a drone equipped with a high-definition camera is used, and the flight route and altitude of the drone are planned by using GIS, and the frequency of farmland inspection is planned, including the following steps:

[0081] According to the performance parameters of the drone, including endurance, flight speed, and the type of sensors carried, set the flight altitude, flight speed, and shooting angle of the drone in the GIS software;

[0082] Use GIS to generate the optimal flight route covering the entire farmland area based on the shape and size of the farmland; set multiple take-off and landing points based on the drone's flight time and shooting requirements; perform simulated flight in GIS to check whether the flight route is reasonable and whether there are any safety hazards or areas that cannot be covered;

[0083] The inspection frequency of farmland is planned according to the month, including: once a week from March to May; twice a week from June to August; twice a week from September to November; once a month from December to February.

[0084] Specifically, obtain the performance parameters of the drone, including the endurance, flight speed, and the type of sensors carried. According to the geographical information of the farmland, including shape, size, and geographical location. According to the performance parameters of the drone and the type of sensor, set the appropriate flight altitude, flight speed, and shooting angle in the GIS software, and determine it according to the actual situation of the farmland. Use GIS software to generate the optimal flight route covering the entire farmland area according to the shape and size of the farmland. According to the endurance of the drone and the shooting requirements, set multiple take-off and landing points to ensure that the flight distance between each take-off and landing point does not exceed the endurance of the drone. When generating the flight route, avoid repeated shooting areas to ensure that each area is only shot once. Perform simulated flight in the GIS software to check whether the flight route is reasonable. Ensure that there are no safety hazards in the flight route, such as avoiding obstacles such as high-voltage lines and buildings. Confirm that the flight route can cover the entire farmland area without missing any areas. Perform drone flight missions according to the planned flight route and frequency. Monitor the flight status of the drone to ensure that it flies according to the planned route. Collect the data captured by the drone for analysis and processing. Use drones to inspect farmland and plan the frequency of farmland inspections according to the month.

[0085] In one embodiment of the present invention, the pre-processed video data, geographic spatial data and environmental parameters are imported into a GIS platform for spatial overlay analysis, including the following steps:

[0086] S1: converting the preprocessed video data, geospatial data and environmental parameters into formats, including: converting video data into vector or raster data through feature extraction, converting geospatial data into vector layers with coordinate information, and converting environmental parameters into shape or surface data;

[0087] S2: Based on the geospatial data, the vector layer of the farmland’s geographic information is imported into the GIS; the data of the monitoring environmental parameters of soil temperature and humidity, air temperature and humidity, light intensity, and carbon dioxide concentration are added as attribute information to the corresponding geospatial data layer, and the video data of the farmland is imported into the GIS;

[0088] S3: Set the data to a unified coordinate system and projection for spatial overlay analysis;

[0089] S4: Analyze the monitored environmental parameters and farmland boundaries, determine the farmland area to which the monitoring point belongs, and perform point-to-polygon overlay, including: importing point data and polygon data into GIS, wherein the point data and polygon data represent the environmental parameters and farmland boundaries detected by the monitoring point, respectively; using the spatial connection or point-in-polygon function of GIS software, overlay and analyze the point data and polygon data, and generate a farmland area layer with environmental parameter information;

[0090] S5: Statistical analysis is performed on the farmland area layer with environmental parameter information superimposed on the point data and polygon data, and the field statistics function of GIS is used to calculate the x-th i The mean and variance of each environmental parameter value in the planting area of ​​​​the crops; according to the analysis results, different thresholds for different environmental parameters are set, and the selection function of the GIS software is used to screen out the farmland areas where the environmental parameters do not meet the standards according to the set thresholds, and mark the screened farmland areas with substandard environmental parameters on the map; increase the number of inspections according to the farmland areas where the environmental parameters do not meet the standards, and increase the inspection frequency planned by month from one inspection per week or month.

[0091] Specifically, the preprocessed video data is converted into vector or raster data through feature extraction. The geospatial data is converted into a vector layer with coordinate information. The environmental parameters are converted into point or surface data. According to the geospatial data, the vector layer of the geographic information of the farmland is imported into the GIS. The data of the monitored environmental parameters are added as attribute information to the corresponding geospatial data layer. The video data of the farmland is imported into the GIS for subsequent analysis. All data are set to a unified coordinate system and projection to ensure the accuracy and consistency of the data. Before performing spatial overlay analysis, ensure that all data layers are in the same coordinate system. Analyze the monitored environmental parameters and farmland boundaries, determine the farmland area to which each point belongs, and perform point and polygon overlay. Use the spatial statistical tools of GIS to perform statistical analysis on the overlay results. Calculate the statistical indicators of the mean and variance to evaluate the growth status of crops. According to the results of the statistical analysis, set different thresholds for different environmental parameters. Soil temperature below 10℃ or above 35℃ is not up to standard, soil moisture below 30% or above 70% is not up to standard, air humidity below 40% or above 60% is not up to standard, air temperature below 10℃ or above 35℃ is not up to standard, light intensity below 200μmol / m 2 / s is not up to standard, and carbon dioxide concentration below 350ppm or above 450ppm is not up to standard. Use the selection function of GIS software to screen out farmland areas with substandard environmental parameters according to the set threshold, and mark the screened farmland areas with substandard environmental parameters on the map; increase the number of inspections according to the farmland areas with substandard environmental parameters, and increase the inspection frequency from weekly or monthly inspections by one in the monthly planned inspection frequency.

[0092] In one embodiment of the present invention, a deep learning algorithm is used to identify pests and diseases in the processed video data, and the xth i Planting crops with pests and diseases, including the following steps:

[0093] Based on the pre-processed video data, the collected images are labeled to identify normal crops and crops with pests and diseases;

[0094] Use the convolutional neural network deep learning model to identify pests and diseases, and input the labeled pest and disease image data into the model for training;

[0095] The xth batch of farmland is inspected based on the video data collected in real time by the GIS-based automatic video inspection system. i Analyze crops and identify normal crops and crops with pests and diseases.

[0096] Specifically, based on the preprocessed video data, the collected images are manually annotated to distinguish normal crops from crops with pests and diseases. During the annotation process, professional image annotation tools, including LabelImg or CVAT, can be used to improve efficiency and accuracy. Establish a convolutional neural network model for pest and disease identification. Input the annotated pest and disease image data into the convolutional neural network model for training. Use the cross entropy loss function and optimizer for model training. Deploy the trained convolutional neural network model into the GIS-based automatic video inspection system. Collect video data of farmland in real time and input it into the convolutional neural network model for analysis. Identify normal crops and crops with pests and diseases, and display the results in real time on the GIS map.

[0097] In one embodiment of the present invention, the growth rate and coverage rate of pest-infested crops are calculated based on the analysis results, and the pest-infested crop variation rate is calculated through comprehensive analysis, including the following steps:

[0098] According to the image recognition results, a patrol is carried out at time t1, and the xth i The number of pixels of pest-infested crops and normal crops of the crops; at the time t2 of the next inspection, the xth pixel identified is counted i The number of pixels of crops with pests and diseases and normal crops;

[0099] By analyzing the formula:

[0100] Get the pixel area M; where q represents the total number of pixels in the image, s x represents the area of ​​the xth pixel, x∈{1,2,...,q}; by substituting into the formula, we can get the xth pixel at time t1 and time t2 respectively. i The pixel area of ​​pest-infested crops and normal crops;

[0101] By analyzing the formula:

[0102] Get the growth rate of pest-infested crops H B (t); where Indicates that at the xth iThe pixel area of ​​the pest-infested crops at the t1th moment of planting crops, Indicates that at the xth i The pixel area of ​​the pest-infested crops at time t2 of planting crops;

[0103] By analyzing the formula:

[0104]

[0105]

[0106] Get the coverage rate of pests and diseases crops F B ;in, Indicates that at the xth i The pixel area of ​​normal crops at the t1th moment of planting crops, Indicates that at the xth i The pixel area of ​​normal crops at the t2th moment, N t1 represents the total pixel area of ​​crops planted at time t1, represents the total pixel area of ​​crops planted at time t2;

[0107] By analyzing the formula:

[0108] Get the pest and disease crop variation rate B, where H B (t) represents the growth rate of pest-infested crops, F B It indicates the coverage rate of crops with pests and diseases.

[0109] Specifically, according to the image analysis results, the pixel area of ​​normal crops and pest crops is obtained by multiplying the number of pixels of normal crops and pest crops in the image by the area of ​​each pixel. By monitoring pests and diseases at different inspection points, the development trend of pests and diseases over time can be understood. The growth rate of pest crops is calculated by counting the pixel area of ​​pest crops at time t1 and the pixel area of ​​pest crops at time t2, which can represent the growth ratio of pest crops within the inspection interval. By calculating the coverage rate of pest crops, the xth inspection interval can be obtained. i The proportion of pest-infested crops among crops; calculating the pest-infested crop variation rate is to combine the growth rate and coverage rate of pest-infested crops to obtain a comprehensive indicator, which reflects the spread speed and scope of pests and diseases in crops.

[0110] In one embodiment of the present invention, analyzing and calculating the growth index of crops according to the collected and processed environmental parameters includes the following steps:

[0111] According to the collected and processed environmental parameters, the environmental parameter matrix (Y ab )n×m ;

[0112] By analyzing the formula:

[0113] Get the standard environmental parameter value Y ab ′ , where a represents the xth i Planting crops, Y b represents the environmental parameter value of the bth type, Y ab Indicates the xth i The environmental parameter value of the bth crop;

[0114] By calculating the contribution of environmental parameters:

[0115] By analyzing the formula:

[0116] Get the influence coefficient K ab ′ , where G ab represents the contribution value of environmental parameters, Y ab ′ Indicates standard environmental parameter values;

[0117] By analyzing the formula:

[0118] Get the growth index of crops δ, where K ab ′ represents the influence coefficient, Y ab ′ Indicates standard environmental parameter values.

[0119] Specifically, an environmental parameter matrix is ​​established based on the collected and processed environmental parameters. This matrix contains parameter values ​​of different types of crops under different environments. An analytical formula is used to calculate the standard environmental parameter value for each crop. This process may involve comparing the differences in the responses of different crops to the same environmental parameter to determine a standard environmental parameter value. Another analytical formula is used to calculate the specific contribution value of each environmental parameter to crop growth. This helps to understand which environmental factors have a greater impact on crop growth. The relationship between the environmental parameter contribution value and the standard environmental parameter value is further analyzed to obtain an influence coefficient. This coefficient reflects the importance or influence of a specific environmental parameter on crop growth. Finally, combined with the previously obtained influence coefficient and the standard environmental parameter value, a comprehensive analytical formula is used to calculate the crop growth index, which can represent the growth status of the crop in the current environment.

[0120] In one embodiment of the present invention, the xth iThe soil contamination rate of the crop planting area includes the following steps:

[0121] Based on the collected and processed video data of the farmland, an image of the soil area is extracted, the RGB value of the soil color is extracted from the image, and the average value of the RGB component of each pixel in the image is calculated;

[0122] By analyzing the formula:

[0123] The soil pollution rate R is obtained, where Red avg 、Green avg 、Blue avg are the average values ​​of red, green and blue respectively, c represents the coefficient of change of soil contamination rate, and λ represents the expected value of soil contamination rate.

[0124] Specifically, from the farmland video data, image processing techniques, including threshold processing, color segmentation, etc., are used to extract images of soil areas. The RGB value of soil color is extracted; for the extracted soil area image, the RGB value of each pixel is calculated. The RGB values ​​of all pixels in the image are averaged to obtain the average values ​​of red, green, and blue. c represents a constant of 0.3, which is used to adjust the change in the contamination rate; the expected value λ of the soil contamination rate is 0.15. First, a certain amount of historical data needs to be collected, including soil RGB values ​​at different time points and known contamination rates. These data will be used to analyze and establish the relationship between RGB values ​​and contamination rates. Statistical analysis methods are used to analyze the correlation between RGB average values ​​and contamination rates. This can help understand the degree of influence of different color components on the contamination rate. According to the results of the correlation analysis, if the RGB value is strongly correlated with the contamination rate, c can be set to a relatively large value to amplify the impact of RGB changes on the contamination rate; if the correlation is weak, c can be set to a smaller value.

[0125] In one embodiment of the present invention, the farmland is comprehensively rated according to the pest and crop variation rate, growth index and soil pollution rate of the farmland, and the farmland inspection is carried out according to the results, including the following steps:

[0126] By analyzing the formula: β=B*ε1+δ*ε2+R*ε3

[0127] The comprehensive rating value of farmland β is obtained; among them, B represents the pest-infested crop variability, δ represents the growth index, and R represents the soil pollution rate; ε1, ε2, and ε3 represent the weights of the pest-infested crop variability, growth index, and soil pollution rate, respectively;

[0128] According to the comprehensive rating value of farmland, the farmland grade area is divided into high-rated value area and low-rated value area. When it is a low-rated value area, the number of inspections is increased based on the inspection frequency planned according to the month, from one inspection per week or month. Otherwise, regular inspections are carried out according to the monthly plan.

[0129] Specifically, ε1, ε2 and ε3 represent the weights of pest and disease crop variability, growth index and soil pollution rate, which are 0.4, 0.4 and 0.2 respectively; according to the calculated farmland comprehensive rating value, the farmland can be divided into high rating value area and low rating value area. According to the farmland comprehensive rating value, the threshold is set to 0.5, and the farmland grade area is divided into high rating value area and low rating value area. When the farmland comprehensive rating value exceeds the threshold and is a low rating value area, the number of inspections is increased on the basis of the inspection frequency planned according to the month, from one inspection per week or month; when the farmland comprehensive rating value is lower than the threshold and is a high rating value area, regular inspections are carried out according to the monthly plan.

[0130] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An automatic video inspection system based on GIS, characterized in that: Includes the following modules: Data collection module: collects geospatial data, environmental data and video data of farmland; Data processing module: pre-processing the collected video data, including: denoising and image enhancement; pre-processing the collected geospatial data and environmental parameters, including: cleaning and normalization; Farmland division module: According to different crop growth requirements, the farmland is divided into n types of crop planting areas, and the i-th planting area is marked Types of crops grown, including ; Data analysis module: import the pre-processed video data, geospatial data and environmental parameters into the GIS platform for spatial overlay analysis; use deep learning algorithms to identify pests and diseases on the processed video data, analyze and identify the first Calculate the growth rate and coverage rate of pest-infested crops according to the analysis results and calculate the pest-infested crop variation rate through comprehensive analysis; Analyze and calculate the growth index of crops according to the collected and processed environmental parameters; Analyze the first the soil contamination rate in the areas where the crops are grown; Farmland rating and inspection module: Based on the pest and crop variation rate, growth index and soil pollution rate of the farmland, the farmland is comprehensively rated and the number of farmland inspections is dynamically adjusted according to the results; The spatial overlay analysis includes: statistical analysis of the farmland area layer with environmental parameter information based on the overlay of point data and polygon data, and using the field statistics function of GIS to calculate the The mean and variance of each environmental parameter value in the planting area of ​​crops; according to the analysis results, different thresholds for different environmental parameters are set, and the selection function of GIS software is used to screen out farmland areas where environmental parameters do not meet the standards according to the set thresholds, and mark the screened farmland areas where environmental parameters do not meet the standards on the map; according to the farmland areas where environmental parameters do not meet the standards, the number of inspections is increased, and the inspection frequency planned by month is increased from one inspection per week or month; Among them, the farmland is comprehensively rated according to the pest and crop variability, growth index and soil pollution rate of the farmland, and the number of farmland inspections is dynamically adjusted according to the results, including the following steps: By analyzing the formula: Get the comprehensive rating of farmland ;in, Indicates the crop pest and disease variation rate, represents the growth index, represents the soil contamination rate; , and represent the weights of crop pest and disease variability, growth index, and soil contamination rate, respectively; According to the comprehensive rating value of farmland, the farmland grade area is divided into high-rated value area and low-rated value area. When it is a low-rated value area, the number of inspections is increased based on the inspection frequency planned according to the month, from one inspection per week or month. Otherwise, regular inspections are carried out according to the monthly plan.

2. The GIS-based automatic video inspection system according to claim 1, characterized in that: Collecting geospatial data, environmental data, and video data of farmland includes the following steps: The geospatial data collection uses a global positioning system (GPS) device to obtain geographical information of the farmland, including: longitude and latitude coordinates, farmland boundaries, farmland area and farmland shape; Environmental sensors are deployed in farmland, including soil temperature and humidity sensors, air temperature and humidity sensors, light sensors, and carbon dioxide concentration sensors, to monitor the environmental parameters of farmland in real time, including soil temperature and humidity, air temperature and humidity, light intensity, and carbon dioxide concentration; Use drones equipped with high-definition cameras, plan the flight route and altitude of the drones through GIS, and plan the frequency of farmland inspections.

3. The GIS-based automatic video inspection system according to claim 2 is characterized in that: Use a drone equipped with a high-definition camera, plan the flight route and altitude of the drone by using GIS, and plan the inspection frequency, including the following steps: According to the performance parameters of the drone, including endurance, flight speed, and the type of sensors carried, set the flight altitude, flight speed, and shooting angle of the drone in the GIS software; Use GIS to generate the optimal flight route covering the entire farmland area based on the shape and size of the farmland; set multiple take-off and landing points based on the drone's flight time and shooting requirements; perform simulated flight in GIS to check whether the flight route is reasonable and whether there are any safety hazards or areas that cannot be covered; The inspection frequency of farmland is planned according to the month, including: once a week from March to May; twice a week from June to August; twice a week from September to November; once a month from December to February.

4. The GIS-based automatic video inspection system according to claim 1, characterized in that: Import the preprocessed video data, geospatial data and environmental parameters into the GIS platform for spatial overlay analysis, including the following steps: S1: converting the preprocessed video data, geospatial data and environmental parameters into formats, including: converting video data into vector or raster data through feature extraction, converting geospatial data into vector layers with coordinate information, and converting environmental parameters into shape or surface data; S2: Based on the geospatial data, the vector layer of the farmland’s geographic information is imported into the GIS; the data of the monitoring environmental parameters of soil temperature and humidity, air temperature and humidity, light intensity, and carbon dioxide concentration are added as attribute information to the corresponding geospatial data layer, and the video data of the farmland is imported into the GIS; S3: Set the data to a unified coordinate system and projection for spatial overlay analysis; S4: Analyze the monitored environmental parameters and farmland boundaries, determine the farmland area to which the monitoring point belongs, and perform point and polygon overlay, including: importing point data and polygon data into GIS, where the point data and polygon data represent the environmental parameters and farmland boundaries detected by the monitoring point, respectively; use the spatial connection or point-in-polygon function of GIS software to overlay and analyze the point data and polygon data to generate a farmland area layer with environmental parameter information.

5. The GIS-based automatic video inspection system according to claim 1 is characterized in that: Use deep learning algorithms to identify pests and diseases in processed video data, analyze and identify the Planting crops with pests and diseases, including the following steps: Based on the pre-processed video data, the collected images are labeled to identify normal crops and crops with pests and diseases; Use the convolutional neural network deep learning model to identify pests and diseases, and input the labeled pest and disease image data into the model for training; The first Analyze crops and identify normal crops and crops with pests and diseases.

6. The GIS-based automatic video inspection system according to claim 5, characterized in that: The growth rate and coverage rate of pest-infested crops are calculated based on the analysis results, and the pest-infested crop variation rate is calculated through comprehensive analysis, including the following steps: According to the image recognition results, Perform an inspection at any time and statistically identify the The number of pixels of pest-infested crops and normal crops; The first The number of pixels of crops with pests and diseases and normal crops; By analyzing the formula: Get pixel area ; where q represents the total number of pixels in the image, represents the area of ​​the xth pixel, x ; By substituting into the formula, we can get Moment and The moment The pixel area of ​​pest-infested crops and normal crops; By analyzing the formula: Get the growth rate of crops with pests and diseases ;in, Indicated in Planting crops The pixel area of ​​pest-infested crops at each moment, Indicated in Planting crops The pixel area of ​​pest-infested crops at the moment; By analyzing the formula: Get the coverage rate of crops with pests and diseases ;in, Indicated in Planting crops The pixel area of ​​normal crops at any moment, Indicated in Planting crops The pixel area of ​​normal crops at any moment, Indicated in The total pixel area of ​​crops planted at the moment, Indicated in The total pixel area of ​​crops planted at the moment; By analyzing the formula: Get the pest and disease crop variation rate ,in, Indicates the growth rate of crops affected by pests and diseases, It indicates the coverage rate of crops with pests and diseases.

7. The GIS-based automatic video inspection system according to claim 1 is characterized in that: According to the collected and processed environmental parameters, the growth index of crops is analyzed and calculated, including the following steps: According to the collected and processed environmental parameters, establish the environmental parameter matrix ; By analyzing the formula: Get standard environmental parameter values ,in Indicates Planting crops, Indicates The environmental parameter values ​​of the species, Indicates Planting crops The environmental parameter values ​​of the species; By calculating the contribution of environmental parameters: By analyzing the formula: Get the influence coefficient ,in represents the contribution value of environmental parameters, Indicates standard environmental parameter values; By analyzing the formula: Get the growth index of crops ,in represents the influence coefficient, Indicates standard environmental parameter values.

8. The GIS-based automatic video inspection system according to claim 1 is characterized in that: According to the collected and processed farmland video data, the The soil contamination rate of the crop planting area includes the following steps: Based on the collected and processed video data of the farmland, an image of the soil area is extracted, the RGB value of the soil color is extracted from the image, and the average value of the RGB component of each pixel in the image is calculated; By analyzing the formula: Get soil contamination rate ,in, , , are the average values ​​of red, green, and blue, respectively. represents the coefficient of change of soil contamination rate, and λ represents the expected value of soil contamination rate.

Images