Water area environment monitoring method and system based on image recognition

Abstract

The invention relates to the technical field of image recognition, in particular to a water area environment monitoring method and system based on image recognition. Firstly, an environment image and a remote sensing image collected by an unmanned aerial vehicle are obtained; secondly, inputting the environment image into an external feature recognition model for recognition and evaluation to obtain a first external environment monitoring value; the external environment influence degree of the water area is evaluated by using external environment monitoring data, so that the accuracy and reliability of water area environment monitoring are improved; then, inputting the remote sensing image into a water area feature recognition model for recognition and evaluation to obtain a second water area environment monitoring value; the influence degree of the water area characteristics is evaluated by using the water area environment monitoring data, so that the accuracy and reliability of water area environment monitoring are improved; and finally, comparing the comprehensive water area environment monitoring value with a water area environment monitoring threshold value, if the comprehensive water area environment monitoring value exceeds the threshold value, outputting early warning prompt information, otherwise, generating water flow monitoring information according to a data result of the water area environment prediction model.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and specifically to a method and system for monitoring water environment based on image recognition. Background Art

[0002] With the acceleration of the industrialization and urbanization processes, the water areas around cities are inevitably polluted. In order to protect the water ecological environment, it is essential to monitor the water environment.

[0003] There are still obvious deficiencies in the prior art for water environment monitoring. On the one hand, the prior art generally uses the image data collected by cameras to evaluate the water environment, and can only monitor the clearly visible entity information, unable to obtain the microscopic information in the water area, which will reduce the accuracy and reliability of water environment monitoring; on the other hand, the prior art does not consider the impact of the surrounding external environment on the water area, and incomplete consideration factors will reduce the accuracy and reliability of water environment monitoring.

[0004] Therefore, a method and system for monitoring water environment based on image recognition are proposed. Summary of the Invention

[0005] The purpose of the present invention is to provide a method and system for monitoring water environment based on image recognition. First, obtain the environmental image and remote sensing image collected by the unmanned aerial vehicle; secondly, input the environmental image into the external feature recognition model for recognition and evaluation to obtain the first external environment monitoring value; use the external environment monitoring data to evaluate the external influence degree of the water area to improve the accuracy and reliability of water environment monitoring; then, input the remote sensing image into the water area feature recognition model for recognition and evaluation to obtain the second water environment monitoring value; use the water environment monitoring data to evaluate the internal influence degree of the water area to improve the accuracy and reliability of water environment monitoring; finally, compare the comprehensive water environment monitoring value with the water environment monitoring threshold. If it exceeds the threshold, output a warning prompt message, otherwise generate water flow monitoring information according to the data result of the water area environmental prediction model.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for monitoring water environment based on image recognition includes:

[0008] Obtain the environmental image of the unmanned aerial vehicle equipped with a camera and the remote sensing image of the unmanned aerial vehicle equipped with a remote sensing sensor;

[0009] Use the external feature recognition model to process the environmental image to obtain external environment monitoring data;

[0010] Comprehensively evaluate the external environment monitoring data and calculate the first external environment monitoring value;

[0011] Process the remote sensing image using the water area feature recognition model to obtain water area environment monitoring data;

[0012] Comprehensively evaluate the water area environment monitoring data and calculate the second water area environment monitoring value;

[0013] Perform weighted calculation on the first external environment monitoring value and the second water area environment monitoring value to obtain the comprehensive water area environment monitoring value;

[0014] Compare the comprehensive water area environment monitoring value with the water area environment monitoring threshold. If the comprehensive water area environment monitoring value exceeds the water area environment monitoring threshold, generate a water area environment warning message and inform the relevant staff; if it is lower than the water area environment monitoring threshold, input the external environment monitoring data and the water area environment monitoring data into the water area regional environment prediction model, and generate water flow monitoring information according to the data results of the water area regional environment prediction model.

[0015] Furthermore, obtain the environmental image captured by the camera carried by the unmanned aerial vehicle and the remote sensing image captured by the remote sensing sensor carried by the unmanned aerial vehicle;

[0016] Furthermore, process the environmental image using the external feature recognition model to obtain external environment monitoring data;

[0017] Furthermore, comprehensively evaluate the external environment monitoring data and calculate the first external environment monitoring value; among them, the specific implementation process of the first external environment monitoring value includes:

[0018] Obtain the environmental image captured by the camera carried by the unmanned aerial vehicle;

[0019] Furthermore, input the environmental image into the external feature recognition model for recognition to obtain external environment monitoring data; among them, the external environment monitoring data includes: urban distribution data, factory distribution data, soil distribution data around the water area, and vegetation distribution data around the water area;

[0020] Furthermore, evaluate each item of distribution data in the external environment monitoring data to obtain multiple evaluation scores;

[0021] Furthermore, perform comprehensive evaluation according to the multiple evaluation scores to obtain the first external environment monitoring value; among them, the calculation formula of the first external environment monitoring value is:

[0022]

[0023] Among them, WBJC represents the first external environment monitoring value; M represents the quantity of external environment monitoring data; among them, α1 represents the urban distribution coefficient; csd i represents the urban distribution data of the i-th monitoring data; csd th represents the set urban distribution threshold; α2 represents the factory distribution coefficient; gsd i represents the factory distribution data of the i-th monitoring data; gsd th represents the set factory distribution threshold; α3 represents the soil distribution coefficient around the water area; trd i represents the soil distribution data around the water area of the i-th monitoring data; trd th represents the set soil distribution threshold around the water area; α4 represents the vegetation distribution data around the water area; zbd i represents the vegetation distribution data around the water area of the i-th monitoring data; zbd th represents the set vegetation distribution threshold around the water area.

[0024] Furthermore, the remote sensing image is processed by using a water area feature recognition model to obtain water area environment monitoring data;

[0025] Furthermore, a comprehensive evaluation is performed on the water area environment monitoring data, and a second water area environment monitoring value is calculated; among them, the specific implementation process of the second water area environment monitoring value includes:

[0026] The remote sensing image is input into the water area feature recognition model for recognition to obtain water area environment monitoring data; among them, the water area environment monitoring data includes color features, shape features, floating object features, and plankton distribution features;

[0027] Furthermore, each feature in the output of the water area feature recognition model is evaluated to obtain multiple evaluation scores;

[0028] Furthermore, a comprehensive evaluation is performed according to the multiple evaluation scores to obtain the second water area environment monitoring value; among them, the calculation formula of the second water area environment monitoring value is:

[0029]

[0030] Among them, SYJC represents the second water area environment monitoring value; N represents the number of monitoring data; β1 represents the color feature coefficient; ys m represents the color transparency of the m-th monitoring data; β2 represents the shape feature coefficient; xzcd m represents the water area length of the m-th monitoring data; xzkd mThe water width represented by the m-th monitoring data; β3 represents the floating object characteristic coefficient; pf m The maximum number of floating objects represented by the m-th monitoring data; pfcd (m,j) The length of the j-th floating object in the m-th monitoring data; pfkd (m,j) The width of the j-th floating object in the m-th monitoring data; β4 represents the plankton distribution characteristic coefficient; fymj m The plankton distribution area represented by the m-th monitoring data; fysd m The plankton distribution depth represented by the m-th monitoring data.

[0031] Further, the first external environment monitoring value and the second water area environment monitoring value are weighted and calculated to obtain a comprehensive water area environment monitoring value; wherein, the calculation formula of the comprehensive water area environment monitoring value is:

[0032] ZHJC = ω * WBJC + (1 - ω) * SYJC;

[0033] Wherein, ZHJC represents the comprehensive water area environment monitoring value; ω represents the weight value assigned to the comprehensive water area environment monitoring value; WBJC represents the first external environment monitoring value; SYJC represents the second water area environment monitoring value.

[0034] Further, the comprehensive water area environment monitoring value is compared with the water area environment monitoring threshold. If the comprehensive water area environment monitoring value exceeds the water area environment monitoring threshold, a water area environment warning message is generated and informed to relevant staff; if it is lower than the water area environment monitoring threshold, the external environment monitoring data and the water area environment monitoring data are input into the water area regional environment prediction model, and water flow monitoring information is generated according to the data result of the water area regional environment prediction model; wherein, the specific implementation process of generating the water flow monitoring information includes:

[0035] The water area is divided into multiple regions;

[0036] Further, the real-time water environment information in the water area environment monitoring data of each region is obtained;

[0037] Further, the real-time soil environment information and real-time vegetation environment information in the external environment monitoring data of each region are obtained;

[0038] Further, the historical water environment information, historical soil environment information and historical vegetation environment information in the historical monitoring database are input into the water area regional environment prediction model for training to obtain environment prediction parameters;

[0039] Further, input the environmental prediction parameters, the real-time water environment information, the real-time soil environment information, and the real-time vegetation environment information into the water area environmental prediction model to update the model parameters and output the real-time environmental parameters. Among them, the real-time environmental parameters include real-time water environment parameters, real-time soil environment parameters, and real-time vegetation environment parameters.

[0040] Further, comprehensively evaluate the real-time environmental information and the real-time environmental parameters to obtain the water flow monitoring information of the water area. The calculation formula of the water flow monitoring information is:

[0041]

[0042] Among them, SLJC represents the water flow monitoring information; P represents the number of divisions of the water area; η (i,1) represents the real-time water environment parameter of the i-th area; INF (i,s) represents the real-time water environment information of the i-th area; η (i,2) represents the real-time soil environment parameter of the i-th area; INF (i,t) represents the real-time soil environment information of the i-th area; η (i,3) represents the real-time vegetation environment parameter of the i-th area; INF (i,z) represents the real-time vegetation environment information of the i-th area.

[0043] A water area environment monitoring system based on image recognition includes: a device control module, a data acquisition module, a first monitoring module, a second monitoring module, a third monitoring module, and a monitoring output module. Among them, the device control module is used to control the startup, pause, and stop of the device; the data acquisition module is used to obtain the image data collected by the drone; the first monitoring module is used to identify and process the environmental image to obtain the first external environmental monitoring value; the second monitoring module is used to identify and process the remote sensing image to obtain the second water area environmental monitoring value; the third monitoring module is used to evaluate and calculate the first external environmental monitoring value and the second water area environmental monitoring value to obtain the comprehensive water area environmental monitoring value; the monitoring output module is used to output the monitoring results and provide monitoring feedback. Among them, the monitoring output module includes a judgment unit, a monitoring warning prompt unit, and a water flow monitoring unit.

[0044] Among them, the calculation formula for the first monitoring module to identify and process the environmental image to obtain the first external environmental monitoring value is:

[0045]

[0046] Among them, WBJC represents the first external environment monitoring value; M represents the quantity of external environment monitoring data; among them, α1 represents the urban distribution coefficient; csd i represents the urban distribution data of the i-th monitoring data; csd th represents the set urban distribution threshold; α2 represents the factory distribution coefficient; gsd i represents the factory distribution data of the i-th monitoring data; gsd th represents the set factory distribution threshold; α3 represents the soil distribution coefficient around the water area; trd i represents the soil distribution data around the water area of the i-th monitoring data; trd th represents the set soil distribution threshold around the water area; α4 represents the vegetation distribution data around the water area; zbd i represents the vegetation distribution data around the water area of the i-th monitoring data; zbd th represents the set vegetation distribution threshold around the water area.

[0047] The second monitoring module is used to identify and process the remote sensing image, and the calculation formula for obtaining the second water area environment monitoring value is:

[0048]

[0049] Among them, SYJC represents the second water area environment monitoring value; N represents the number of monitoring data; β1 represents the color feature coefficient; ys m represents the color transparency of the m-th monitoring data; β2 represents the shape feature coefficient; xzcd m represents the water area length of the m-th monitoring data; xzkd m represents the water area width of the m-th monitoring data; β3 represents the floating object feature coefficient; pf m represents the maximum number of floating objects of the m-th monitoring data; pfcd (m,j) represents the length of the j-th floating object of the m-th monitoring data; pfkd (m,j) represents the width of the j-th floating object of the m-th monitoring data; β4 represents the plankton distribution feature coefficient; fymj m represents the plankton distribution area of the m-th monitoring data; fysd m represents the plankton distribution depth of the m-th monitoring data.

[0050] The third monitoring module is used to evaluate and calculate the first external environment monitoring value and the second water area environment monitoring value, and the calculation formula for obtaining the comprehensive water area environment monitoring value is:

[0051] ZHJC = ω * WBJC + (1 - ω) * SYJC;

[0052] Among them, ZHJC represents the comprehensive water environment monitoring value; ω represents the weight value assigned to the comprehensive water environment monitoring value; WBJC represents the first external environment monitoring value; SYJC represents the second water environment monitoring value.

[0053] The calculation formula for the water flow monitoring unit to obtain water flow monitoring information is:

[0054]

[0055] Among them, SLJC represents the water flow monitoring information; P represents the number of water area divisions; η (i,1) represents the real-time water environment parameter of the i-th area; INF (i,s) represents the real-time water environment information of the i-th area; η (i,2) represents the real-time soil environment parameter of the i-th area; INF (i,t) represents the real-time soil environment information of the i-th area; η (i,3) represents the real-time vegetation environment parameter of the i-th area; INF (i,z) represents the real-time vegetation environment information of the i-th area.

[0056] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0057] 1. The present invention proposes an external environment monitoring function for monitoring the external environmental impact on water areas; this function identifies and processes environmental images through an external feature recognition model to obtain external environment monitoring data; the external environment monitoring data records the distribution of cities, factories, surrounding soil, and vegetation; this function evaluates and calculates different external environment distribution data to obtain the first external environment monitoring value; the first external environment monitoring value reflects the degree of influence of the external environment on the water environment; this function can effectively improve the accuracy and reliability of water environment monitoring.

[0058] 2. The present invention proposes a water environment monitoring function for monitoring the internal impact on water areas; this function identifies and processes remote sensing images through a water area feature recognition model to obtain water environment monitoring data; the water environment monitoring data records the color transparency, shape changes, number of floating objects, and plankton distribution of the water area; this function evaluates and calculates different water environment feature data to obtain the second water environment monitoring value; the second water environment monitoring value reflects the degree of internal influence of the water environment; this function can effectively improve the accuracy and reliability of water environment monitoring.

[0059] 3. The present invention proposes a water flow monitoring function for water areas to monitor the impact of the real-time environment on water flow; this function uses historical environmental data to train an environmental prediction model for water areas to obtain environmental prediction parameters; uses the environmental prediction parameters and real-time environmental data to update the model parameters to obtain real-time environmental parameters; this function evaluates the real-time environmental information and the real-time environmental parameters to obtain water flow monitoring information for water areas; this water flow monitoring information reflects the degree of impact of the real-time environment on water flow; this function can effectively improve the accuracy and reliability of water area environmental monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a schematic flowchart of the method for monitoring water area environment based on image recognition of the present invention;

[0061] Figure 2 It is a schematic structural diagram of the water area feature recognition model of the present invention;

[0062] Figure 3 It is a schematic structural diagram of the water area environment monitoring system based on image recognition of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0064] With the acceleration of the industrialization and urbanization processes, the surrounding water areas are inevitably polluted. In order to protect the water area ecological environment, it is essential to monitor the water area environment.

[0065] Water area environmental monitoring refers to the process of observing and evaluating biological and non-biological elements in water areas to monitor water quality, the health status of the ecosystem, or environmental changes. This kind of monitoring helps to protect water resources, maintain the water ecological balance, prevent water pollution and ecosystem degradation and other problems.

[0066] The existing technology still has obvious deficiencies in water area environmental monitoring. On the one hand, the existing technology generally uses the image data collected by cameras to evaluate the water area environment, and can only monitor the clearly visible entity information, and cannot obtain the microscopic information in the water area, which will reduce the accuracy and reliability of water area environmental monitoring; on the other hand, the existing technology does not consider the impact of the surrounding external environment on the water area, and the incomplete consideration of factors will reduce the accuracy and reliability of water area environmental monitoring.

[0067] Embodiment 1

[0068] In the embodiment of the present application, the specific implementation process will be realized by a water area environment monitoring method based on image recognition. Refer to Figure 1 for the process description of the method proposed in this embodiment; the water area environment monitoring method based on image recognition includes:

[0069] S10. Obtain the environmental image captured by the UAV camera and the remote sensing image collected by the remote sensing sensor;

[0070] S20. Input the environmental image and the remote sensing image into an external feature recognition model and a water area feature recognition model respectively to obtain external environmental monitoring data and water area environmental monitoring data;

[0071] S30. Conduct a comprehensive evaluation on the external environmental monitoring data and the water area environmental monitoring data, and calculate and obtain a first external environmental monitoring value and a second water area environmental monitoring value respectively;

[0072] S40. Perform a weighted calculation on the first external environmental monitoring value and the second water area environmental monitoring value to obtain a comprehensive water area environmental monitoring value;

[0073] S50. Compare the comprehensive water area environmental monitoring value with the water area environmental monitoring threshold. If the comprehensive water area environmental monitoring value exceeds the water area environmental monitoring threshold, generate a water area environmental warning message and notify relevant staff; if it is lower than the water area environmental monitoring threshold, input the external environmental monitoring data and the water area environmental monitoring data into the water area regional environment prediction model, and generate water flow monitoring information according to the data result of the water area regional environment prediction model.

[0074] Furthermore, the specific implementation process of a water area environment monitoring method based on image recognition is as follows:

[0075] Obtain the environmental image of the UAV-mounted camera and the remote sensing image of the UAV-mounted remote sensing sensor;

[0076] Use an external feature recognition model to process the environmental image to obtain external environmental monitoring data;

[0077] Conduct a comprehensive evaluation on the external environmental monitoring data and calculate and obtain a first external environmental monitoring value;

[0078] Use a water area feature recognition model to process the remote sensing image to obtain water area environmental monitoring data;

[0079] Conduct a comprehensive evaluation on the water area environmental monitoring data and calculate and obtain a second water area environmental monitoring value;

[0080] Perform weighted calculation on the first external environment monitoring value and the second water area environment monitoring value to obtain a comprehensive water area environment monitoring value;

[0081] Compare the comprehensive water area environment monitoring value with the water area environment monitoring threshold. If the comprehensive water area environment monitoring value

[0082] exceeds the water area environment monitoring threshold, generate a water area environment warning message and inform the relevant staff; if it is lower than the water area environment monitoring threshold, input the external environment monitoring data and the water area environment monitoring data into the water area regional environment prediction model, and generate water flow monitoring information according to the data result of the water area regional environment prediction model.

[0083] In this embodiment, a water area environment monitoring method based on image recognition is proposed; First, obtain the environmental image and remote sensing image collected by the unmanned aerial vehicle; Secondly, input the environmental image into the external feature recognition model for recognition and evaluation to obtain the first external environment monitoring value; Use the external environment monitoring data to evaluate the external environmental impact degree of the water area to improve the accuracy and reliability of water area environment monitoring; Then, input the remote sensing image into the water area feature recognition model for recognition and evaluation to obtain the second water area environment monitoring value; Use the water area environment monitoring data to evaluate the influence degree of the water area features to improve the accuracy and reliability of water area environment monitoring; Finally, compare the comprehensive water area environment monitoring value with the water area environment monitoring threshold. If it exceeds the threshold, output a warning prompt message, otherwise generate water flow monitoring information according to the data result of the water area regional environment prediction model to improve the accuracy and reliability of water area environment monitoring.

[0084] For specific illustration, it is elaborated in combination with the following embodiments as follows:

[0085] Obtain the environmental image of the camera carried by the unmanned aerial vehicle and the remote sensing image of the remote sensing sensor carried by the unmanned aerial vehicle;

[0086] The specific process of obtaining the remote sensing image in this embodiment includes: collecting electromagnetic wave signals such as visible light, infrared rays, and microwaves, and converting the processed signals into digital signals; Processing and analyzing the data, extracting and identifying the characteristic information required for monitoring, and generating a remote sensing image.

[0087] Furthermore, use the external feature recognition model to process the environmental image to obtain external environment monitoring data; Among them, the external feature recognition model includes: an input layer, a multi-scale feature extraction layer, a marking layer, an identification layer, and an output layer; The specific implementation process of this model includes:

[0088] Input the environmental image into the input layer to obtain environmental features;

[0089] Further, input the environmental features into the multi-scale feature extraction layer to obtain multi-scale environmental features; wherein, the multi-scale feature extraction layer is implemented by convolution operations of different sizes.

[0090] Further, input the multi-scale environmental features into the marking layer to obtain the anchor boxes of the environmental features; wherein, the anchor boxes are set to five different sizes.

[0091] Further, extract the features in the anchor boxes and input them into the recognition layer to obtain the environmental recognition result.

[0092] Further, input the environmental recognition result into the output layer to obtain the external environmental monitoring data.

[0093] In this embodiment, an external feature recognition model is used to extract external environmental monitoring data. The external feature recognition model is a deep model based on a convolutional neural network. This model takes the environmental images collected by the drone as input, and outputs the external environmental data affecting the water area after feature extraction, feature marking, and feature recognition operations; the method of image recognition can improve the efficiency and accuracy of water area environmental monitoring.

[0094] Further, comprehensively evaluate the external environmental monitoring data to calculate the first external environmental monitoring value; wherein, the specific implementation process of the first external environmental monitoring value includes:

[0095] Obtain the environmental images of the camera carried by the drone.

[0096] Further, input the environmental images into the external feature recognition model for recognition to obtain the external environmental monitoring data; wherein, the external environmental monitoring data includes: urban distribution data, factory distribution data, soil distribution data around the water area, and vegetation distribution data around the water area.

[0097] Further, evaluate each distribution data in the external environmental monitoring data to obtain multiple evaluation scores.

[0098] Further, perform a comprehensive evaluation based on the multiple evaluation scores to obtain the first external environmental monitoring value; wherein, the calculation formula of the first external environmental monitoring value is:

[0099]

[0100] α1 + α2 + α3 + α4 = 1.0;

[0101] wherein, WBJC represents the first external environmental monitoring value; M represents the number of external environmental monitoring data; wherein, α1 represents the urban distribution coefficient; csd iThe urban distribution data represented as the i-th monitoring data; csd th The set urban distribution threshold; α2 represents the factory distribution coefficient; gsd i The factory distribution data represented as the i-th monitoring data; gsd th The set factory distribution threshold; α3 represents the soil distribution coefficient around the water area; trd i The soil distribution data around the water area represented as the i-th monitoring data; trd th The set soil distribution threshold around the water area; α4 represents the vegetation distribution coefficient around the water area; zbd i The vegetation distribution data around the water area represented as the i-th monitoring data; zbd th The set vegetation distribution threshold around the water area.

[0102] The urban distribution threshold, factory distribution threshold, soil distribution threshold around the water area, and vegetation distribution threshold around the water area in this embodiment may vary with different location positions and ecologies. Therefore, the set values need to be flexibly set by relevant personnel according to the actual situation.

[0103] The distribution coefficients in this embodiment represent the influence weights of different external environmental characteristics on the water area environment. The larger the coefficient, the higher the influence degree. In this embodiment, the urban distribution coefficient, factory distribution coefficient, soil distribution coefficient around the water area, and vegetation distribution coefficient around the water area are respectively set to 0.3, 0.3, 0.2, and 0.2. Of course, the distribution coefficients can also be adjusted according to the actual situation.

[0104] This embodiment proposes an external environment monitoring function for monitoring the external environmental impact on the water area; this function uses an external feature recognition model to identify and process the environmental image to obtain external environment monitoring data; the external environment monitoring data records the distribution of cities, factories, surrounding soil, and vegetation; this function calculates and evaluates different external environment distribution data to obtain the first external environment monitoring value; the first external environment monitoring value reflects the influence degree of the external environment on the water area environment; this function can effectively improve the accuracy and reliability of water area environment monitoring.

[0105] Furthermore, use a water area feature recognition model to process the remote sensing image to obtain water area environment monitoring data;

[0106] Among them, the structure of the water area feature recognition model is as Figure 2 shown, including: an input layer, a water area feature extraction layer, a water area feature marking layer, a water area feature recognition layer, and an output layer; the specific implementation process of this model includes:

[0107] Input the remote sensing image into the input layer to obtain remote sensing features;

[0108] Further, input the remote sensing features into the water area feature extraction layer to obtain water area environment features;

[0109] Further, input the water area environment features into the water area feature marking layer to obtain the anchor boxes of the water area environment features; wherein, the anchor boxes are set to five different sizes;

[0110] Further, extract the features in the anchor boxes and input them into the water area feature recognition layer to obtain the water area environment recognition result;

[0111] Further, input the water area environment recognition result into the output layer to obtain the water area environment monitoring data.

[0112] Further, comprehensively evaluate the water area environment monitoring data and calculate to obtain a second water area environment monitoring value; wherein, the specific implementation process of the second water area environment monitoring value includes:

[0113] Input the remote sensing image into the water area feature recognition model for recognition to obtain the water area environment monitoring data; wherein, the water area environment monitoring data includes color features, shape features, floating object features, and plankton distribution features;

[0114] In this embodiment, the shape of the water area is approximated to a rectangle, which is to facilitate the subsequent evaluation and calculation of the shape features in the output of the water area feature recognition model.

[0115] In this embodiment, the color features include the degree of color transparency; the shape features include the width and length of the water area; the floating object features include the number of floating objects, the length and width of the floating objects; the plankton distribution features include the plankton distribution area and the plankton distribution depth.

[0116] Further, evaluate each feature in the output of the water area feature recognition model to obtain multiple evaluation scores;

[0117] Further, perform a comprehensive evaluation based on the multiple evaluation scores to obtain the second water area environment monitoring value; wherein, the calculation formula of the second water area environment monitoring value is:

[0118]

[0119] β1 + β2 + β3 + β4 = 1.0;

[0120] Wherein, SYJC represents the second water area environment monitoring value; N represents the number of monitoring data; β1 represents the color feature coefficient; ys mThe color transparency of the m-th monitoring data; β2 represents the shape feature coefficient; xzcd m The water area length of the m-th monitoring data; xzkd m The water area width of the m-th monitoring data; β3 represents the floating object feature coefficient; pf m The maximum number of floating objects of the m-th monitoring data; pfcd (m,j) The length of the j-th floating object of the m-th monitoring data; pfkd (m,j) The width of the j-th floating object of the m-th monitoring data; β4 represents the plankton distribution feature coefficient; fymj m The plankton distribution area of the m-th monitoring data; fysd m The plankton distribution depth of the m-th monitoring data.

[0121] The feature coefficients in this embodiment represent the influence weights of different water area characteristics on the water area environment. The larger the coefficient, the higher the influence degree. In this embodiment, the color feature coefficient, shape feature coefficient, floating object feature coefficient, and plankton distribution feature coefficient are all set to 0.25. Of course, the feature coefficients can also be adjusted according to the actual situation.

[0122] In this embodiment, a water area environment monitoring function is proposed to monitor the internal influence situation of the water area environment; this function performs recognition processing on the remote sensing image through the water area feature recognition model to obtain the water area environment monitoring data; the water area environment monitoring data records the color transparency, shape change, floating object quantity, and plankton distribution of the water area; this function obtains the second water area environment monitoring value through the evaluation and calculation of different water area environment feature data; the second water area environment monitoring value reflects the internal influence degree of the water area environment; this function can effectively improve the accuracy and reliability of water area environment monitoring.

[0123] Furthermore, the first external environment monitoring value and the second water area environment monitoring value are weighted and calculated to obtain the comprehensive water area environment monitoring value; where, the calculation formula of the comprehensive water area environment monitoring value is:

[0124] ZHJC = ω * WBJC + (1 - ω) * SYJC;

[0125] Where, ZHJC represents the comprehensive water area environment monitoring value; ω represents the weight value assigned to the comprehensive water area environment monitoring value; WBJC represents the first external environment monitoring value; SYJC represents the second water area environment monitoring value. ω in this embodiment is set to 0.5;

[0126] In this embodiment, the comprehensive situation of the water area environment is reflected by combining external environment monitoring and water area environment monitoring. The comprehensive water area environment monitoring value calculated using the first external environment monitoring value and the second water area environment monitoring value is used to comprehensively monitor the water area regional environment, thereby effectively improving the accuracy and reliability of water area environment monitoring.

[0127] Further, the comprehensive water area environment monitoring value is compared with the water area environment monitoring threshold. If the comprehensive water area environment monitoring value exceeds the water area environment monitoring threshold, a water area environment warning message is generated and informed to relevant staff; wherein, the water area environment monitoring threshold can be flexibly adjusted according to the actual situation; if it is lower than the water area environment monitoring threshold, the external environment monitoring data and the water area environment monitoring data are input into the water area regional environment prediction model, and water flow monitoring information is generated according to the data result of the water area regional environment prediction model.

[0128] Among them, the specific implementation process of generating the water flow monitoring information includes:

[0129] The water area is divided into multiple regions.

[0130] Further, the real-time water environment information in the water area environment monitoring data of each region is obtained.

[0131] Further, the real-time soil environment information and real-time vegetation environment information in the external environment monitoring data of each region are obtained.

[0132] Further, the historical water environment information, historical soil environment information, and historical vegetation environment information in the historical monitoring database are input into the water area regional environment prediction model for training to obtain environment prediction parameters.

[0133] Further, the environment prediction parameters, the real-time water environment information, the real-time soil environment information, and the real-time vegetation environment information are input into the water area regional environment prediction model to update the model parameters, and real-time environment parameters are output; wherein, the real-time environment parameters include: real-time water environment parameters, real-time soil environment parameters, and real-time vegetation environment parameters; wherein, the water area regional environment prediction model adopts an LSTM network structure.

[0134] Further, the real-time environment information and the real-time environment parameters are comprehensively evaluated to obtain water area regional water flow monitoring information; the calculation formula of the water flow monitoring information is:

[0135]

[0136] Among them, SLJC represents the water flow monitoring information; P represents the number of water area regions divided; η (i,1)The real-time water environment parameters represented as the i-th area; INF (i,s) The real-time water environment information represented as the i-th area; η (i,2) The real-time soil environment parameters represented as the i-th area; INF (i,t) The real-time soil environment information represented as the i-th area; η (i,3) The real-time vegetation environment parameters represented as the i-th area; INF (i,z) The real-time vegetation environment information represented as the i-th area.

[0137] In this embodiment, a water flow monitoring function for a water area is proposed to monitor the influence of the real-time environment on the water flow; this function uses historical environment data to train an environmental prediction model for the water area environment to obtain environmental prediction parameters; uses the environmental prediction parameters and real-time environment data to update the model parameters to obtain real-time environment parameters; this function evaluates the real-time environment information and the real-time environment parameters to obtain water area water flow monitoring information; this water flow monitoring information reflects the degree of influence of the real-time environment on the water flow; this function can effectively improve the accuracy and reliability of water area environment monitoring.

[0138] Embodiment 2

[0139] As an implementation manner of the present invention, referring to Figure 3 , a water area environment monitoring system based on image recognition includes: a device control module, a data acquisition module, a first monitoring module, a second monitoring module, a third monitoring module, and a monitoring output module;

[0140] Among them, the device control module is used to control the start, pause, and stop of the device;

[0141] The data acquisition module is used to obtain image data collected by a drone;

[0142] The first monitoring module is used to perform recognition processing on the environmental image to obtain a first external environment monitoring value;

[0143] The second monitoring module is used to perform recognition processing on the remote sensing image to obtain a second water area environment monitoring value;

[0144] The third monitoring module is used to perform evaluation calculations on the first external environment monitoring value and the second water area environment monitoring value to obtain a comprehensive water area environment monitoring value;

[0145] The monitoring output module is used to output the monitoring result and provide monitoring feedback; among them, the monitoring output module includes a judgment unit, a monitoring warning prompt unit, and a water flow monitoring unit.

[0146] Among them, the first monitoring module is used to identify and process environmental images, and the calculation formula for obtaining the first external environmental monitoring value is:

[0147]

[0148] Among them, WBJC represents the first external environmental monitoring value; M represents the quantity of external environmental monitoring data; among them, α1 represents the urban distribution coefficient; csd i represents the urban distribution data of the i-th monitoring data; csd th represents the set urban distribution threshold; α2 represents the factory distribution coefficient; gsd i represents the factory distribution data of the i-th monitoring data; gsd th represents the set factory distribution threshold; α3 represents the soil distribution coefficient around water areas; trd i represents the soil distribution data around water areas of the i-th monitoring data; trd th represents the set soil distribution threshold around water areas; α4 represents the vegetation distribution data around water areas; zbd i represents the vegetation distribution data around water areas of the i-th monitoring data; zbd th represents the set vegetation distribution threshold around water areas.

[0149] The second monitoring module is used to identify and process remote sensing images, and the calculation formula for obtaining the second water area environmental monitoring value is:

[0150]

[0151] Among them, SYJC represents the second water area environmental monitoring value; N represents the number of monitoring data; β1 represents the color feature coefficient; ys m represents the color transparency of the m-th monitoring data; β2 represents the shape feature coefficient; xzcd m represents the water area length of the m-th monitoring data; xzkd m represents the water area width of the m-th monitoring data; β3 represents the floating object feature coefficient; pf m represents the maximum number of floating objects of the m-th monitoring data; pfcd (m,j) represents the length of the j-th floating object of the m-th monitoring data; pfkd (m,j) represents the width of the j-th floating object of the m-th monitoring data; β4 represents the plankton distribution feature coefficient; fymj m represents the plankton distribution area of the m-th monitoring data; fysd m represents the plankton distribution depth of the m-th monitoring data.

[0152] The third monitoring module is used to evaluate and calculate the first external environment monitoring value and the second water area environment monitoring value. The calculation formula for obtaining the comprehensive water area environment monitoring value is as follows:

[0153] ZHJC = ω * WBJC + (1 - ω) * SYJC;

[0154] Among them, ZHJC represents the comprehensive water area environment monitoring value; ω represents the weight value assigned to the comprehensive water area environment monitoring value; WBJC represents the first external environment monitoring value; SYJC represents the second water area environment monitoring value.

[0155] The calculation formula for the water flow monitoring unit to obtain water flow monitoring information is as follows:

[0156]

[0157] Among them, SLJC represents water flow monitoring information; P represents the number of water area divisions; η (i,1) represents the real-time water environment parameter of the i-th area; INF (i,s) represents the real-time water environment information of the i-th area; η (i,2) represents the real-time soil environment parameter of the i-th area; INF (i,t) represents the real-time soil environment information of the i-th area; η (i,3) represents the real-time vegetation environment parameter of the i-th area; INF (i,z) represents the real-time vegetation environment information of the i-th area.

[0158] In this embodiment, a water area environment monitoring system based on image recognition is proposed. The system includes a device control module, a data acquisition module, a first monitoring module, a first monitoring module, a third monitoring module, and a monitoring output module. Among them, the data acquisition module acquires the environmental images captured by the camera and the remote sensing images generated by the remote sensing sensor. The first monitoring module, the first monitoring module, and the third monitoring module are used to monitor the surrounding environment and the water area body of the water area to obtain different monitoring values. The monitoring values reflect the influence degree of the external environment characteristics and water area characteristics on the change of the water area environment. The monitoring output module compares the threshold through the monitoring warning prompt unit to judge whether to output a warning message. If it is lower than the set threshold, the water flow monitoring unit is used to output water flow monitoring information, which can improve the accuracy and reliability of the water area environment monitoring.

[0159] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A water environment monitoring method based on image recognition, characterized in that Including: Obtain the environmental image captured by the camera on the drone and the remote sensing image captured by the remote sensing sensor on the drone; Process the environmental image using an external feature recognition model to obtain external environmental monitoring data; Comprehensively evaluate the external environmental monitoring data and calculate the first external environmental monitoring value; Process the remote sensing image using a water area feature recognition model to obtain water area environmental monitoring data; Comprehensively evaluate the water area environmental monitoring data and calculate the second water area environmental monitoring value; Perform weighted calculation on the first external environmental monitoring value and the second water area environmental monitoring value to obtain a comprehensive water area environmental monitoring value; Compare the comprehensive water area environmental monitoring value with the water area environmental monitoring threshold. If the comprehensive water area environmental monitoring value exceeds the water area environmental monitoring threshold, generate a water area environmental warning message and inform relevant staff; if it is lower than the water area environmental monitoring threshold, input the external environmental monitoring data and the water area environmental monitoring data into the water area regional environmental prediction model, and generate water flow monitoring information based on the data results of the water area regional environmental prediction model.

2. The water environment monitoring method based on image recognition according to claim 1, characterized in that The specific process of processing the environmental image using an external feature recognition model to obtain external environmental monitoring data includes: Obtain the environmental image captured by the camera on the drone; Input the environmental image into the external feature recognition model for recognition to obtain external environmental monitoring data; among them, the external environmental monitoring data includes: urban distribution data, factory distribution data, soil distribution data around the water area, and vegetation distribution data around the water area.

3. The method for monitoring water environment based on image recognition according to claim 1, wherein The specific implementation process of comprehensively evaluating the external environmental monitoring data and calculating the first external environmental monitoring value includes: Obtain the urban distribution data, factory distribution data, soil distribution data around the water area, and vegetation distribution data around the water area in the external environmental monitoring data; Evaluate each distribution data in the external environmental monitoring data to obtain multiple evaluation scores; Perform comprehensive evaluation based on the multiple evaluation scores to obtain the first external environmental monitoring value.

4. The water environment monitoring method based on image recognition according to claim 1, characterized in that The specific implementation process of comprehensively evaluating the water area environmental monitoring data and calculating the second water area environmental monitoring value includes: Input the remote sensing image into the water area feature recognition model for recognition to obtain water area environmental monitoring data; among them, the water area environmental monitoring data includes color features, shape features, floating object features, and plankton distribution features; Evaluate each feature in the output of the water area feature recognition model to obtain multiple evaluation scores; Perform comprehensive evaluation based on the multiple evaluation scores to obtain the second water area environmental monitoring value.

5. A method for monitoring water environment based on image recognition according to claim 1, characterized in that, The specific implementation process of inputting the external environmental monitoring data and the water area environmental monitoring data into the water area regional environmental prediction model and generating water flow monitoring information based on the data results of the water area regional environmental prediction model includes: Divide the water area region into multiple regions; Obtain the real-time water environment information in the water area environmental monitoring data of each region; Obtain the real-time soil environment information and real-time vegetation environment information in the external environmental monitoring data of each region; Input the historical water environment information, historical soil environment information, and historical vegetation environment information in the historical monitoring database into the water area environmental prediction model for training to obtain environmental prediction parameters; Input the environmental prediction parameters, the real-time water environment information, the real-time soil environment information, and the real-time vegetation environment information into the water area environmental prediction model to update the model parameters and output real-time environmental parameters; wherein, the real-time environmental parameters include: real-time water environment parameters, real-time soil environment parameters, and real-time vegetation environment parameters; Comprehensively evaluate the real-time environmental information and the real-time environmental parameters to obtain water flow monitoring information for the water area.

6. An image recognition-based water environment monitoring system, characterized in that, Include: An equipment control module, a data acquisition module, a first monitoring module, a second monitoring module, a third monitoring module, and a monitoring output module; wherein, the equipment control module is used to control the start, pause, and stop of the equipment; the data acquisition module is used to obtain image data collected by the drone; the first monitoring module is used to identify and process the environmental image to obtain a first external environmental monitoring value; the second monitoring module is used to identify and process the remote sensing image to obtain a second water area environmental monitoring value; the third monitoring module is used to evaluate and calculate the first external environmental monitoring value and the second water area environmental monitoring value to obtain a comprehensive water area environmental monitoring value; the monitoring output module is used to output the monitoring result and provide monitoring feedback; wherein, the monitoring output module includes a judgment unit, a monitoring warning prompt unit, and a water flow monitoring unit.

Images