Method and system for detecting safety hazards through artificial intelligence and remote telemetry image recognition technology

Abstract

The present application relates to the field of image processing technology, and provides a method and system for realizing safety hazard inspection through artificial intelligence and remote telemetry image recognition technology. The method includes obtaining multi-source image information corresponding to each target area in the park; obtaining meteorological data information within a preset time period of the park; for each target area, obtaining the safety hazard prediction model, historical safety hazard information, environmental sensitivity information and hazard degree information corresponding to the target area, and inputting the multi-source image information corresponding to the target area into the safety hazard prediction model to obtain multiple predicted safety hazards corresponding to the target area, and scoring each predicted safety hazard based on the meteorological data information, historical safety hazard information, environmental sensitivity information and hazard degree information; and prioritizing each predicted safety hazard based on the score corresponding to each predicted safety hazard to obtain the safety hazard inspection result. The method improves the accuracy of the inspection results.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method and system for realizing safety hazard inspection through artificial intelligence and remote telemetry image recognition technology. Background Art

[0002] With the increasing complexity of infrastructure in modern industrial parks, the identification and management of safety hazards have become crucial. However, the existing safety hazard inspection methods face the problem of too single inspection results in practical applications, and cannot provide comprehensive hazard information to staff. This limitation leads to missed or misdetected hazards, which in turn affects the effectiveness of safety management and makes it difficult to help staff effectively detect safety hazards. Summary of the invention

[0003] The present application provides a method and system for implementing safety hazard inspection through artificial intelligence and remote telemetry image recognition technology to solve the problems raised by the above-mentioned background technology.

[0004] In a first aspect, the present application provides a method for implementing safety hazard inspection by using artificial intelligence and remote telemetry image recognition technology, including:

[0005] Acquire multi-source image information corresponding to each target area in the park through a preset telemetry image acquisition module in the park; wherein the multi-source image information includes visible light images, hyperspectral images and thermal imaging images;

[0006] Obtain the park's meteorological data information within a preset time period in the future;

[0007] For each of the target areas, a safety hazard prediction model, historical safety hazard information, environmental sensitivity information, and hazard level information corresponding to the target area are obtained from a database, and multi-source image information corresponding to the target area is input into the safety hazard prediction model to obtain a plurality of predicted safety hazards corresponding to the target area, and each predicted safety hazard is scored based on the meteorological data information, the historical safety hazard information, the environmental sensitivity information, and the hazard level information;

[0008] The predicted safety hazards are prioritized based on the scores corresponding to the predicted safety hazards, and the safety hazard inspection results in the park are obtained.

[0009] In a possible implementation, inputting the multi-source image information corresponding to the target area into the safety hazard prediction model to obtain a plurality of predicted safety hazards corresponding to the target area includes:

[0010] Inputting the multi-source image information into the safety hazard prediction model; wherein the safety hazard prediction model includes an image type recognition layer, a visible light image feature extraction layer, a hyperspectral image feature extraction layer, a thermal imaging image feature extraction layer, a feature fusion layer and a fully connected layer;

[0011] The image type recognition layer performs type recognition on each image of the multi-source image information respectively to obtain a type recognition result of each image, and inputs each image into a corresponding feature extraction layer based on the type recognition result corresponding to each image;

[0012] The visible light image feature extraction layer performs feature extraction on the received visible light image to obtain a first feature vector;

[0013] The hyperspectral image feature extraction layer extracts features from the received hyperspectral image to obtain a second feature vector;

[0014] The thermal imaging image feature extraction layer extracts features from the received thermal imaging image to obtain a third feature vector;

[0015] The feature fusion layer fuses the first feature vector, the second feature vector, and the third feature vector to obtain a fused feature vector;

[0016] The fully connected layer performs a linear transformation on the fused feature vector based on a preset linear transformation rule to obtain a linear transformation result corresponding to the fused feature vector, and maps the linear transformation result through a preset Sigmoid function to obtain multiple predicted safety hazards.

[0017] In a possible implementation, the meteorological data information includes a plurality of environmental parameter change curves.

[0018] Scoring each of the predicted safety hazards based on the meteorological data information, the historical safety hazard information, the environmental sensitivity information, and the hazard level information includes:

[0019] For each of the environmental parameter change curves, determining the environmental parameter level corresponding to the environmental parameter change curve based on a preset environmental parameter level classification rule;

[0020] For each of the predicted safety hazards, determining a first hazard score corresponding to the predicted safety hazard based on the environmental parameter level corresponding to each of the environmental parameter change curves and the environmental sensitive information;

[0021] For each of the predicted safety hazards, determining a second hazard score corresponding to the predicted safety hazard based on the hazard level information;

[0022] For each of the predicted safety hazards, determining a third hazard score corresponding to the predicted safety hazard based on the historical safety hazard information;

[0023] For each of the predicted safety hazards, a first hazard score, a second hazard score, and a third hazard score corresponding to the predicted safety hazard are weighted based on a preset weight setting rule to obtain a score corresponding to the predicted safety hazard.

[0024] In a possible implementation, the environmental sensitive information includes hazard values ​​corresponding to various safety hazards at different environmental parameter levels of various environmental parameters, and determining the first hazard score corresponding to the predicted safety hazard based on the environmental parameter levels corresponding to the environmental parameter change curves and the environmental sensitive information includes:

[0025] Determining in the environmental sensitive information the target hazard value corresponding to the predicted safety hazard at the environmental parameter level corresponding to each environmental parameter change curve;

[0026] The target hidden danger values ​​are added together to obtain the first hidden danger score.

[0027] In a possible implementation, determining the third hazard score corresponding to the predicted safety hazard based on the historical safety hazard information includes:

[0028] Determining whether the predicted safety hazard has occurred based on the historical safety hazard information;

[0029] If not, determining that the third hidden danger score is 0;

[0030] If so, the probability of the predicted safety hazard occurring is determined based on the historical safety hazard information, and the probability is multiplied by a preset score to obtain the third hazard score.

[0031] In a possible implementation, the method further includes:

[0032] For each of the target areas, binding the target area with multi-source image information corresponding to the target area to obtain intermediate binding information;

[0033] Each of the intermediate binding information is bound with the potential safety hazard inspection result to obtain target binding information, and the target binding information is stored in a database.

[0034] In a possible implementation manner, the step of binding the target area with multi-source image information corresponding to the target area to obtain intermediate binding information includes:

[0035] Obtaining the area code corresponding to the target area;

[0036] Performing feature encoding on the multi-source image information based on a preset image information encoding rule to obtain an image encoding sequence corresponding to the image information;

[0037] The image coding sequence is placed after the region coding to obtain the intermediate binding information.

[0038] In a possible implementation, the step of binding each of the intermediate binding information with the potential safety hazard inspection result to obtain target binding information includes:

[0039] Arrange the intermediate binding information in order from top to bottom based on a preset sorting rule to obtain a first initial matrix;

[0040] Performing matrix coding on the safety hazard inspection result based on a preset matrix coding rule to obtain a second initial matrix; wherein the first initial matrix and the second initial matrix do not have the same characters;

[0041] Determine whether the first initial matrix and the second initial matrix are matrices of the same type;

[0042] If not, compare the number of rows of the first initial matrix with the number of rows of the second initial matrix, and determine the larger number of rows as the target number of rows, and compare the number of columns of the first initial matrix with the number of columns of the second initial matrix, and determine the larger number of columns as the target number of columns;

[0043] Constructing a first intermediate matrix and a second intermediate matrix; the first intermediate matrix is ​​the same as the second intermediate matrix, the number of rows of the first intermediate matrix is ​​greater than the target number of rows, the number of columns of the first intermediate matrix is ​​greater than the target number of columns, the characters of the first intermediate matrix are the same, and the characters of the first intermediate matrix do not exist in the first initial matrix and the second initial matrix;

[0044] Replacing characters at corresponding positions of the first intermediate matrix with characters of the first initial matrix to obtain a first target matrix, and replacing characters at corresponding positions of the second intermediate matrix with characters of the second initial matrix to obtain a second target matrix;

[0045] The first target matrix and the second target matrix are added to obtain the target binding information.

[0046] In a second aspect, the present application provides a safety hazard inspection system implemented by artificial intelligence and remote telemetry image recognition technology, including:

[0047] The first acquisition module is used to acquire multi-source image information corresponding to each target area in the park through a preset telemetry image acquisition module in the park; wherein the multi-source image information includes visible light images, hyperspectral images and thermal imaging images;

[0048] The second acquisition module is used to obtain the meteorological data information of the park within a preset time period in the future;

[0049] A scoring module is used to obtain, for each of the target areas, a safety hazard prediction model, historical safety hazard information, environmental sensitivity information, and hazard level information corresponding to the target area in a database, input the multi-source image information corresponding to the target area into the safety hazard prediction model, obtain multiple predicted safety hazards corresponding to the target area, and score each of the predicted safety hazards based on the meteorological data information, the historical safety hazard information, the environmental sensitivity information, and the hazard level information;

[0050] The sorting module is used to prioritize the predicted safety hazards based on the scores corresponding to the predicted safety hazards, and obtain the safety hazard inspection results in the park.

[0051] The present application provides a method for implementing safety hazard inspection by artificial intelligence and remote telemetry image recognition technology, wherein the method comprises: obtaining multi-source image information corresponding to each target area in the park through a preset telemetry image acquisition module in the park; wherein the multi-source image information comprises visible light images, hyperspectral images and thermal imaging images; obtaining meteorological data information of the park in a preset time period in the future;

[0052] For each target area, the safety hazard prediction model, historical safety hazard information, environmental sensitivity information and hazard degree information corresponding to the target area are obtained in the database, and the multi-source image information corresponding to the target area is input into the safety hazard prediction model to obtain multiple predicted safety hazards corresponding to the target area, and score each predicted safety hazard based on the meteorological data information, the historical safety hazard information, the environmental sensitivity information and the hazard degree information; and the predicted safety hazards are prioritized based on the scores corresponding to each predicted safety hazard to obtain the safety hazard inspection results in the park. On the one hand, this method can comprehensively obtain a variety of image information of the target area by using visible light images, hyperspectral images and thermal imaging images. These different types of image information complement each other, providing multi-dimensional data from macro to micro, from surface to internal, which is helpful to comprehensively identify and detect possible safety hazards. On the other hand, the scoring method for each predicted safety hazard integrates information from multiple dimensions of meteorological data information, historical safety hazard information, environmental sensitivity information and hazard degree information, and realizes a more accurate risk assessment of each predicted safety hazard. The multi-dimensional scoring method improves the accuracy of the scoring, helps to more scientifically determine the priority order of predicting safety hazards, and avoids the misjudgment that may be caused by traditional single-dimensional scoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.

[0054] Figure 1 A flowchart of a method for implementing a safety hazard inspection method using artificial intelligence and remote telemetry image recognition technology provided in an embodiment of the present application;

[0055] Figure 2 A schematic block diagram of the structure of a safety hazard inspection system implemented by artificial intelligence and remote telemetry image recognition technology provided in an embodiment of the present application;

[0056] Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0057] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0059] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0060] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0061] With the increasing complexity of infrastructure in modern industrial parks, the identification and management of safety hazards have become crucial. However, existing safety hazard inspection methods face the problem of too single inspection results in practical applications, and are unable to provide comprehensive hazard information to staff. This limitation leads to missed or false detections of hidden dangers, which in turn affects the effectiveness of safety management and makes it difficult to help staff effectively troubleshoot safety hazards. To this end, the present application provides a method and system for implementing safety hazard inspection through artificial intelligence and remote telemetry image recognition technology to solve the above problems.

[0062] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0063] See also Figure 1 , Figure 1 A flowchart of a method for implementing a safety hazard inspection method using artificial intelligence and remote telemetry image recognition technology provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, the method for realizing safety hazard inspection through artificial intelligence and remote telemetry image recognition technology provided in the embodiment of the present application includes steps S1 to S4.

[0064] Step S1, obtaining multi-source image information corresponding to each target area in the park through a preset telemetry image acquisition module in the park; wherein the multi-source image information includes visible light images, hyperspectral images and thermal imaging images.

[0065] Step S2: Obtaining meteorological data information of the park within a preset time period in the future.

[0066] Step S3: for each of the target areas, obtain the safety hazard prediction model, historical safety hazard information, environmental sensitivity information and hazard level information corresponding to the target area in the database, and input the multi-source image information corresponding to the target area into the safety hazard prediction model to obtain multiple predicted safety hazards corresponding to the target area, and score each predicted safety hazard based on the meteorological data information, the historical safety hazard information, the environmental sensitivity information and the hazard level information.

[0067] Step S4: prioritize the predicted safety hazards based on the scores corresponding to the predicted safety hazards, and obtain safety hazard inspection results in the park.

[0068] In this embodiment, as described in step S1 above, multi-source image information corresponding to each target area in the park is obtained through a preset telemetry image acquisition module in the park; wherein the multi-source image information includes visible light images, hyperspectral images, and thermal imaging images. Specifically, the telemetry image acquisition module is pre-set in each target area in the park, and the telemetry image acquisition module includes a visible light camera, a hyperspectral camera, and an infrared camera. The telemetry image acquisition module can be fixedly set in each target area in the park, or the telemetry image acquisition module can be mounted on a drone or a satellite, and the telemetry image acquisition module is carried to the shooting position of each target area by the drone or the satellite to obtain multi-source image information of each target area. It can be understood that visible light images can be used to see the appearance, color, and shape of objects such as buildings and equipment in the target area, and can be used to observe defects in objects at the macro level or whether the equipment is operating normally, such as surface cracks, and hyperspectral images are used to detect tiny chemical changes, such as corrosion of buildings or equipment. Thermal imaging images are used to display the temperature distribution of objects, and can be used to detect whether the equipment is operating normally or to detect whether there are stress defects inside the object. Visible light images, hyperspectral images, and thermal imaging images help to achieve comprehensive detection of safety hazards in the target area.

[0069] As described in step S2 above, the meteorological data information of the park in the future preset time period is obtained. Specifically, the meteorological data information is obtained through the official website of the Meteorological Bureau. The meteorological data information includes but is not limited to the changes in temperature, humidity, wind speed and rainfall in the preset time period.

[0070] As described in the above step S3, for each of the target areas, the safety hazard prediction model, historical safety hazard information, environmental sensitivity information and hazard level information corresponding to the target area are obtained in the database, and the multi-source image information corresponding to the target area is input into the safety hazard prediction model to obtain multiple predicted safety hazards corresponding to the target area, and each predicted safety hazard is scored based on the meteorological data information, the historical safety hazard information, the environmental sensitivity information and the hazard level information. Specifically, first, a target text corresponding to the target area is obtained in a database through the area coding of the target area; wherein the target text includes a safety hazard prediction model, historical safety hazard information, environmentally sensitive information and hazard level information, the safety hazard prediction model is a pre-trained deep learning model, the environmentally sensitive information includes the hazard values ​​corresponding to each safety hazard under different environmental parameter levels of each environmental parameter, the hazard level information includes the score of the hazard caused by each safety hazard, the environmentally sensitive information and the hazard level information are both obtained through experiments, and the historical safety hazard information records the hazard type and time of each safety hazard occurring in the target area; then, the multi-source image information corresponding to the target area is input into the safety hazard prediction model to obtain multiple predicted safety hazards corresponding to the target area; finally, each predicted safety hazard is scored based on the meteorological data information, the historical safety hazard information, the environmentally sensitive information and the hazard level information. It can be understood that step 3 scores each predicted safety hazard based on the meteorological data information, the historical safety hazard information, the environmental sensitivity information and the hazard level information, thereby achieving multi-dimensional scoring of the predicted safety hazard, which helps to improve the accuracy of the scoring results and, in turn, improve the effectiveness of the inspection results.

[0071] As described in step S4 above, the predicted safety hazards are prioritized based on the scores corresponding to the predicted safety hazards, and the safety hazard inspection results in the park are obtained. Specifically, the predicted safety hazards corresponding to all target areas are sorted according to the scores, with the predicted safety hazards with high scores being ranked in front and the predicted safety hazards with low scores being ranked in the back. In this way, a list can be obtained to tell the staff which hidden dangers in the park are the most urgent and need to be dealt with first. This list is the final safety hazard inspection result.

[0072] The method provided in this embodiment, on the one hand, can comprehensively obtain a variety of image information of the target area by using visible light images, hyperspectral images and thermal imaging images. These different types of image information complement each other, providing multi-dimensional data from macro to micro, from surface to internal, which helps to comprehensively identify and detect potential safety hazards. On the other hand, the scoring method for each predicted safety hazard integrates information from multiple dimensions such as meteorological data information, historical safety hazard information, environmental sensitivity information and hazard level information, and realizes a more accurate risk assessment of each predicted safety hazard. The multi-dimensional scoring method improves the accuracy of the scoring, helps to more scientifically determine the priority order of predicted safety hazards, and avoids the misjudgment that may be caused by traditional single-dimensional scoring.

[0073] In some embodiments, inputting the multi-source image information corresponding to the target area into the safety hazard prediction model to obtain a plurality of predicted safety hazards corresponding to the target area includes:

[0074] Inputting the multi-source image information into the safety hazard prediction model; wherein the safety hazard prediction model includes an image type recognition layer, a visible light image feature extraction layer, a hyperspectral image feature extraction layer, a thermal imaging image feature extraction layer, a feature fusion layer and a fully connected layer;

[0075] The image type recognition layer performs type recognition on each image of the multi-source image information respectively to obtain a type recognition result of each image, and inputs each image into a corresponding feature extraction layer based on the type recognition result corresponding to each image;

[0076] The visible light image feature extraction layer performs feature extraction on the received visible light image to obtain a first feature vector;

[0077] The hyperspectral image feature extraction layer extracts features from the received hyperspectral image to obtain a second feature vector;

[0078] The thermal imaging image feature extraction layer extracts features from the received thermal imaging image to obtain a third feature vector;

[0079] The feature fusion layer fuses the first feature vector, the second feature vector, and the third feature vector to obtain a fused feature vector;

[0080] The fully connected layer performs a linear transformation on the fused feature vector based on a preset linear transformation rule to obtain a linear transformation result corresponding to the fused feature vector, and maps the linear transformation result through a preset Sigmoid function to obtain multiple predicted safety hazards.

[0081] The method provided in this embodiment enables the safety hazard prediction model to recognize and classify different types of image information by setting an image type recognition layer. The accuracy of image processing is improved, so that each type of image information can enter the corresponding feature extraction layer, ensuring the effectiveness and accuracy of feature extraction. The feature extraction layer of each image is specially designed for that type of image, so that the feature information of that image type can be extracted to the maximum extent, thereby improving the prediction accuracy of the safety hazard prediction model.

[0082] In some embodiments, the meteorological data information includes a plurality of environmental parameter change curves, and the scoring of each of the predicted safety hazards based on the meteorological data information, the historical safety hazard information, the environmental sensitivity information and the hazard level information includes:

[0083] For each of the environmental parameter change curves, determining the environmental parameter level corresponding to the environmental parameter change curve based on a preset environmental parameter level classification rule;

[0084] For each of the predicted safety hazards, a first hazard score corresponding to the predicted safety hazard is determined based on the environmental parameter level corresponding to each of the environmental parameter change curves and the environmental sensitive information; specifically, the environmental sensitive information includes hazard values ​​corresponding to each of the safety hazards at different environmental parameter levels of each environmental parameter. When determining the first hazard score corresponding to the predicted safety hazard based on the environmental parameter level corresponding to each of the environmental parameter change curves and the environmental sensitive information, first, target hazard values ​​corresponding to the predicted safety hazard at the environmental parameter level corresponding to each of the environmental parameter change curves are respectively determined in the environmental sensitive information, and then, each of the target hazard values ​​is added to obtain the first hazard score;

[0085] For each of the predicted safety hazards, determine a second hazard score corresponding to the predicted safety hazard based on the hazard level information; specifically, use the score corresponding to the predicted safety hazard in the hazard level information as the second hazard score;

[0086] For each of the predicted safety hazards, a third hazard score corresponding to the predicted safety hazard is determined based on the historical safety hazard information; specifically, whether the predicted safety hazard has occurred is determined based on the historical safety hazard information; if not, the third safety hazard score is determined to be 0; if so, the probability of the predicted safety hazard occurring is determined based on the historical safety hazard information, and the probability is multiplied by a preset score to obtain the third hazard score; wherein, when determining the probability of the predicted safety hazard occurring based on the historical safety hazard information, first, a target historical safety hazard that is the same as the predicted safety hazard is determined in the historical safety hazard information, then, the number of occurrences of the target historical safety hazard is determined in the historical safety hazard information, and finally, the ratio between the number of occurrences of the target historical safety hazard and the total number of safety hazards occurring in the historical safety hazard information is used as the probability;

[0087] For each of the predicted safety hazards, a first hazard score, a second hazard score, and a third hazard score corresponding to the predicted safety hazard are weighted based on a preset weight setting rule to obtain a score corresponding to the predicted safety hazard. The first hazard score corresponds to a first weight, the second hazard score corresponds to a second weight, and the third hazard score corresponds to a third weight, the first weight is greater than the third weight, and the second weight is greater than the first weight.

[0088] Exemplarily, the meteorological data information includes a temperature change curve, a humidity change curve, a wind speed change curve and a rainfall change curve. The temperature level corresponding to the temperature change curve is medium, the humidity level corresponding to the humidity change curve is high, the wind force level corresponding to the wind speed change curve is high, and the rainfall level corresponding to the rainfall change curve is medium. The predicted safety hazard is a first-level crack in the XX wall. The environmental sensitivity information is shown in Table 1. The score corresponding to the first-level crack in the XX wall in the hazard level information is 80, the probability corresponding to the first-level crack in the XX wall is 0.02, the preset score is 100, the first weight is 0.3, the second weight is 0.6, and the third weight is 0.1. The predicted safety hazard is a score corresponding to the first-level crack in the XX wall of 0.3×(20+70+35+50)+0.6×80+0.1×0.02×100=100.7.

[0089] Table 1

[0090]

[0091] The method provided in this embodiment achieves multi-dimensional scoring of the predicted safety hazards, which helps to improve the accuracy of the scoring results and, in turn, improve the effectiveness of the inspection results.

[0092] In some embodiments, the method further comprises:

[0093] For each of the target areas, binding the target area with multi-source image information corresponding to the target area to obtain intermediate binding information;

[0094] Each of the intermediate binding information is bound with the potential safety hazard inspection result to obtain target binding information, and the target binding information is stored in a database.

[0095] The step of binding the target area with the multi-source image information corresponding to the target area to obtain intermediate binding information includes:

[0096] Obtaining the area code corresponding to the target area;

[0097] Performing feature encoding on the multi-source image information based on a preset image information encoding rule to obtain an image encoding sequence corresponding to the image information;

[0098] The image coding sequence is placed after the region coding to obtain the intermediate binding information.

[0099] The step of binding each of the intermediate binding information with the potential safety hazard inspection result to obtain target binding information includes:

[0100] Arrange the intermediate binding information in order from top to bottom based on a preset sorting rule to obtain a first initial matrix;

[0101] Performing matrix coding on the safety hazard inspection result based on a preset matrix coding rule to obtain a second initial matrix; wherein the first initial matrix and the second initial matrix do not have the same characters;

[0102] Determine whether the first initial matrix and the second initial matrix are matrices of the same type;

[0103] If not, compare the number of rows of the first initial matrix with the number of rows of the second initial matrix, and determine the larger number of rows as the target number of rows, and compare the number of columns of the first initial matrix with the number of columns of the second initial matrix, and determine the larger number of columns as the target number of columns;

[0104] Constructing a first intermediate matrix and a second intermediate matrix; the first intermediate matrix is ​​the same as the second intermediate matrix, the number of rows of the first intermediate matrix is ​​greater than the target number of rows, the number of columns of the first intermediate matrix is ​​greater than the target number of columns, the characters of the first intermediate matrix are the same, and the characters of the first intermediate matrix do not exist in the first initial matrix and the second initial matrix;

[0105] Replacing characters at corresponding positions of the first intermediate matrix with characters of the first initial matrix to obtain a first target matrix, and replacing characters at corresponding positions of the second intermediate matrix with characters of the second initial matrix to obtain a second target matrix;

[0106] The first target matrix and the second target matrix are added to obtain the target binding information.

[0107] The method provided in this embodiment, on the one hand, obtains intermediate binding information by binding the target area with its corresponding multi-source image information, and further binds each intermediate binding information with the safety hazard inspection result, thereby realizing a systematic association between data, so that relevant information can be accurately and systematically stored and managed in the database. This systematic data management method improves the efficiency of data retrieval and provides a solid data foundation for subsequent analysis and decision-making.

[0108] See also Figure 2 , Figure 2 A schematic block diagram of the structure of a safety hazard inspection system implemented by artificial intelligence and remote telemetry image recognition technology provided in an embodiment of the present application, such as Figure 2 As shown, the safety hazard inspection system is realized through artificial intelligence and remote telemetry image recognition technology, including:

[0109] The first acquisition module 110 is used to acquire multi-source image information corresponding to each target area in the park through a preset telemetry image acquisition module in the park; wherein the multi-source image information includes visible light images, hyperspectral images and thermal imaging images.

[0110] The second acquisition module 120 is used to acquire the meteorological data information of the park within a preset time period in the future.

[0111] The scoring module 130 is used to obtain, for each target area, a safety hazard prediction model, historical safety hazard information, environmental sensitivity information and hazard level information corresponding to the target area in the database, and input the multi-source image information corresponding to the target area into the safety hazard prediction model to obtain multiple predicted safety hazards corresponding to the target area, and score each predicted safety hazard based on the meteorological data information, the historical safety hazard information, the environmental sensitivity information and the hazard level information.

[0112] The sorting module 140 is used to prioritize the predicted safety hazards based on the scores corresponding to the predicted safety hazards, and obtain the safety hazard inspection results in the park.

[0113] It should be noted that technical personnel in the relevant technical field can clearly understand that, for the convenience and conciseness of description, the specific working process of the system and each module described above can refer to the corresponding process in the aforementioned embodiment of the safety hazard inspection method through artificial intelligence and remote telemetry image recognition technology, and will not be repeated here.

[0114] The safety hazard inspection system provided by the above embodiment through artificial intelligence and remote telemetry image recognition technology can be implemented in the form of a computer program. The computer program can be used in Figure 3 The system is run on the terminal device 200 shown.

[0115] See also Figure 3 , Figure 3 The terminal device 200 provided in the embodiment of the present application is a schematic block diagram of the structure, wherein the terminal device 200 includes a processor 201 and a memory 202, wherein the processor 201 and the memory 202 are connected via a system bus 203, wherein the memory 202 may include a non-volatile storage medium and an internal memory.

[0116] The non-volatile storage medium can store a computer program. The computer program includes program instructions, and when the program instructions are executed by the processor 201, the processor 201 can execute any of the above-mentioned safety hazard inspection methods using artificial intelligence and remote telemetry image recognition technology.

[0117] The processor 201 is used to provide computing and control capabilities to support the operation of the entire terminal device 200.

[0118] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor 201, the processor 201 can execute any of the above-mentioned safety hazard inspection methods through artificial intelligence and remote telemetry image recognition technology.

[0119] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the terminal device 200 involved in the scheme of the present application. The specific terminal device 200 may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0120] It should be understood that the processor 201 may be a central processing unit (CPU), and the processor 201 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0121] In some embodiments, the processor 201 is used to run a computer program stored in the memory to implement the following steps:

[0122] Acquire multi-source image information corresponding to each target area in the park through a preset telemetry image acquisition module in the park; wherein the multi-source image information includes visible light images, hyperspectral images and thermal imaging images;

[0123] Obtain the park's meteorological data information within a preset time period in the future;

[0124] For each of the target areas, a safety hazard prediction model, historical safety hazard information, environmental sensitivity information, and hazard level information corresponding to the target area are obtained from a database, and multi-source image information corresponding to the target area is input into the safety hazard prediction model to obtain a plurality of predicted safety hazards corresponding to the target area, and each predicted safety hazard is scored based on the meteorological data information, the historical safety hazard information, the environmental sensitivity information, and the hazard level information;

[0125] The predicted safety hazards are prioritized based on the scores corresponding to the predicted safety hazards, and the safety hazard inspection results in the park are obtained.

[0126] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by one or more processors, the one or more processors implement a safety hazard inspection method through artificial intelligence and remote telemetry image recognition technology as provided in an embodiment of the present application.

[0127] The computer-readable storage medium may be an internal storage unit of the terminal device 200 in the aforementioned embodiment, such as a hard disk or memory of the terminal device 200. The computer-readable storage medium may also be an external storage device of the terminal device 200, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped with the terminal device 200.

[0128] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for detecting potential safety hazards by using artificial intelligence and remote telemetry image recognition technology, characterized in that: include: Acquire multi-source image information corresponding to each target area in the park through a preset telemetry image acquisition module in the park; wherein the multi-source image information includes visible light images, hyperspectral images and thermal imaging images; Obtain the park's meteorological data information within a preset time period in the future; For each of the target areas, a safety hazard prediction model, historical safety hazard information, environmental sensitivity information, and hazard level information corresponding to the target area are obtained from a database, and multi-source image information corresponding to the target area is input into the safety hazard prediction model to obtain a plurality of predicted safety hazards corresponding to the target area, and each predicted safety hazard is scored based on the meteorological data information, the historical safety hazard information, the environmental sensitivity information, and the hazard level information; Prioritize each of the predicted safety hazards based on the scores corresponding to each of the predicted safety hazards to obtain safety hazard inspection results in the park; The method further comprises: For each of the target areas, binding the target area with multi-source image information corresponding to the target area to obtain intermediate binding information; Binding each of the intermediate binding information with the potential safety hazard inspection result to obtain target binding information, and storing the target binding information in a database; The step of binding the target area with the multi-source image information corresponding to the target area to obtain intermediate binding information includes: Obtaining the area code corresponding to the target area; Performing feature encoding on the multi-source image information based on a preset image information encoding rule to obtain an image encoding sequence corresponding to the image information; Placing the image coding sequence after the region coding to obtain the intermediate binding information; The step of binding each of the intermediate binding information with the potential safety hazard inspection result to obtain target binding information includes: Arrange the intermediate binding information in order from top to bottom based on a preset sorting rule to obtain a first initial matrix; Performing matrix coding on the safety hazard inspection result based on a preset matrix coding rule to obtain a second initial matrix; wherein the first initial matrix and the second initial matrix do not have the same characters; Determine whether the first initial matrix and the second initial matrix are matrices of the same type; If not, compare the number of rows of the first initial matrix with the number of rows of the second initial matrix, and determine the larger number of rows as the target number of rows, and compare the number of columns of the first initial matrix with the number of columns of the second initial matrix, and determine the larger number of columns as the target number of columns; Constructing a first intermediate matrix and a second intermediate matrix; the first intermediate matrix is ​​the same as the second intermediate matrix, the number of rows of the first intermediate matrix is ​​greater than the target number of rows, the number of columns of the first intermediate matrix is ​​greater than the target number of columns, the characters of the first intermediate matrix are the same, and the characters of the first intermediate matrix do not exist in the first initial matrix and the second initial matrix; Replacing characters at corresponding positions of the first intermediate matrix with characters of the first initial matrix to obtain a first target matrix, and replacing characters at corresponding positions of the second intermediate matrix with characters of the second initial matrix to obtain a second target matrix; The first target matrix and the second target matrix are added to obtain the target binding information.

2. The method for realizing safety hazard inspection by artificial intelligence and remote telemetry image recognition technology according to claim 1 is characterized in that: The step of inputting the multi-source image information corresponding to the target area into the safety hazard prediction model to obtain a plurality of predicted safety hazards corresponding to the target area includes: Inputting the multi-source image information into the safety hazard prediction model; wherein the safety hazard prediction model includes an image type recognition layer, a visible light image feature extraction layer, a hyperspectral image feature extraction layer, a thermal imaging image feature extraction layer, a feature fusion layer and a fully connected layer; The image type recognition layer performs type recognition on each image of the multi-source image information respectively to obtain a type recognition result of each image, and inputs each image into a corresponding feature extraction layer based on the type recognition result corresponding to each image; The visible light image feature extraction layer performs feature extraction on the received visible light image to obtain a first feature vector; The hyperspectral image feature extraction layer extracts features from the received hyperspectral image to obtain a second feature vector; The thermal imaging image feature extraction layer extracts features from the received thermal imaging image to obtain a third feature vector; The feature fusion layer fuses the first feature vector, the second feature vector, and the third feature vector to obtain a fused feature vector; The fully connected layer performs a linear transformation on the fused feature vector based on a preset linear transformation rule to obtain a linear transformation result corresponding to the fused feature vector, and maps the linear transformation result through a preset Sigmoid function to obtain multiple predicted safety hazards.

3. The method for realizing safety hazard inspection by artificial intelligence and remote telemetry image recognition technology according to claim 1 is characterized in that: The meteorological data information includes a plurality of environmental parameter change curves, Scoring each of the predicted safety hazards based on the meteorological data information, the historical safety hazard information, the environmental sensitivity information, and the hazard level information includes: For each of the environmental parameter change curves, determining the environmental parameter level corresponding to the environmental parameter change curve based on a preset environmental parameter level classification rule; For each of the predicted safety hazards, determining a first hazard score corresponding to the predicted safety hazard based on the environmental parameter level corresponding to each of the environmental parameter change curves and the environmental sensitive information; For each of the predicted safety hazards, determining a second hazard score corresponding to the predicted safety hazard based on the hazard level information; For each of the predicted safety hazards, determining a third hazard score corresponding to the predicted safety hazard based on the historical safety hazard information; For each of the predicted safety hazards, a first hazard score, a second hazard score, and a third hazard score corresponding to the predicted safety hazard are weighted based on a preset weight setting rule to obtain a score corresponding to the predicted safety hazard.

4. The method for realizing safety hazard inspection by artificial intelligence and remote telemetry image recognition technology according to claim 3 is characterized in that: The environmental sensitive information includes the hazard values ​​corresponding to each safety hazard at different environmental parameter levels of each environmental parameter, and determining the first hazard score corresponding to the predicted safety hazard based on the environmental parameter levels corresponding to each environmental parameter change curve and the environmental sensitive information includes: Determining in the environmental sensitive information the target hazard value corresponding to the predicted safety hazard at the environmental parameter level corresponding to each environmental parameter change curve; The target hidden danger values ​​are added together to obtain the first hidden danger score.

5. The method for realizing safety hazard inspection by artificial intelligence and remote telemetry image recognition technology according to claim 3 is characterized in that: The determining, based on the historical safety hazard information, a third hazard score corresponding to the predicted safety hazard includes: Determining whether the predicted safety hazard has occurred based on the historical safety hazard information; If not, determining that the third hidden danger score is 0; If so, the probability of the predicted safety hazard occurring is determined based on the historical safety hazard information, and the probability is multiplied by a preset score to obtain the third hazard score.

6. A safety hazard inspection system using artificial intelligence and remote telemetry image recognition technology, characterized in that: include: The first acquisition module is used to acquire multi-source image information corresponding to each target area in the park through a preset telemetry image acquisition module in the park; wherein the multi-source image information includes visible light images, hyperspectral images and thermal imaging images; The second acquisition module is used to obtain the meteorological data information of the park within a preset time period in the future; A scoring module is used to obtain, for each of the target areas, a safety hazard prediction model, historical safety hazard information, environmental sensitivity information, and hazard level information corresponding to the target area in a database, input the multi-source image information corresponding to the target area into the safety hazard prediction model, obtain multiple predicted safety hazards corresponding to the target area, and score each of the predicted safety hazards based on the meteorological data information, the historical safety hazard information, the environmental sensitivity information, and the hazard level information; A sorting module is used to prioritize the predicted safety hazards based on the scores corresponding to the predicted safety hazards, obtain the safety hazard inspection results in the park, and for each target area, bind the target area with the multi-source image information corresponding to the target area to obtain intermediate binding information, and bind each intermediate binding information with the safety hazard inspection result to obtain target binding information, and store the target binding information in a database; The step of binding the target area with the multi-source image information corresponding to the target area to obtain intermediate binding information includes: Obtaining the area code corresponding to the target area; Performing feature encoding on the multi-source image information based on a preset image information encoding rule to obtain an image encoding sequence corresponding to the image information; Placing the image coding sequence after the region coding to obtain the intermediate binding information; The step of binding each of the intermediate binding information with the potential safety hazard inspection result to obtain target binding information includes: Arrange the intermediate binding information in order from top to bottom based on a preset sorting rule to obtain a first initial matrix; Performing matrix coding on the safety hazard inspection result based on a preset matrix coding rule to obtain a second initial matrix; wherein the first initial matrix and the second initial matrix do not have the same characters; Determine whether the first initial matrix and the second initial matrix are matrices of the same type; If not, compare the number of rows of the first initial matrix with the number of rows of the second initial matrix, and determine the larger number of rows as the target number of rows, and compare the number of columns of the first initial matrix with the number of columns of the second initial matrix, and determine the larger number of columns as the target number of columns; Constructing a first intermediate matrix and a second intermediate matrix; the first intermediate matrix is ​​the same as the second intermediate matrix, the number of rows of the first intermediate matrix is ​​greater than the target number of rows, the number of columns of the first intermediate matrix is ​​greater than the target number of columns, the characters of the first intermediate matrix are the same, and the characters of the first intermediate matrix do not exist in the first initial matrix and the second initial matrix; Replacing characters at corresponding positions of the first intermediate matrix with characters of the first initial matrix to obtain a first target matrix, and replacing characters at corresponding positions of the second intermediate matrix with characters of the second initial matrix to obtain a second target matrix; The first target matrix and the second target matrix are added to obtain the target binding information.

Images