Early warning method and device for landslide and debris flow disasters and electronic equipment

Abstract

The invention provides a landslide and debris flow disaster early warning method and device and electronic equipment, and the method comprises the steps: obtaining a terrain elevation data set corresponding to a to-be-recognized target region; each piece of terrain elevation data in the terrain elevation data set comprises a corresponding data point position in the target area; inputting the terrain elevation data set into a preset landslide and debris flow identification model to determine whether at least one piece of target terrain elevation data corresponding to the landslide and debris flow disaster risk exists in the terrain elevation data set; and executing early warning operation based on each piece of target terrain elevation data under the condition of determining that at least one piece of target terrain elevation data exists. Therefore, landslide and debris flow topographic features are mined through a machine learning method, and the identification precision of landslide and debris flow disaster risks is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of natural disaster early warning, and in particular to a landslide and debris flow disaster early warning method, device and electronic equipment. Background Art

[0002] Landslides and mudslides are common natural geological disasters that are highly destructive. Once they occur, they pose a great threat to the safety of nearby facilities and residents.

[0003] For industrial facilities such as oil and gas pipelines and power transmission facilities that are widely located in wild areas, the identification and early warning of landslides and mudslides can greatly ensure the safety of the facilities. For disasters that have already occurred, rapid identification, monitoring and early warning can be achieved, helping relevant personnel to quickly restore facilities in the disaster area and avoid further losses. Summary of the invention

[0004] The present invention provides a landslide and debris flow disaster early warning method, device, electronic equipment and storage medium, which are used to at least solve the defect that the prior art cannot identify and warn landslide and debris flow disasters in wild areas.

[0005] The present invention provides a landslide and debris flow disaster early warning method, the method comprising: obtaining a terrain elevation data set corresponding to a target area to be identified; each terrain elevation data in the terrain elevation data set comprises a corresponding data point position located in the target area; the terrain elevation data set is input into a preset landslide and debris flow identification model to determine whether there is at least one target terrain elevation data corresponding to a landslide and debris flow disaster risk in the terrain elevation data set; each training sample in the sample data set of the landslide and debris flow identification model comprises a terrain elevation data sample and a corresponding label, the label being a landslide and debris flow disaster risk or a landslide and debris flow disaster risk; and when it is determined that the at least one target terrain elevation data exists, performing an early warning operation based on each of the target terrain elevation data.

[0006] According to a landslide and mud-rock flow disaster warning method provided by the present invention, the warning operation is performed based on each of the target terrain elevation data, including: determining the risk area range corresponding to the landslide and mud-rock flow disaster risk based on the data point position corresponding to each of the target terrain elevation data; and performing the risk warning operation based on the risk area range.

[0007] According to a landslide and mudslide disaster warning method provided by the present invention, after determining the risk area range corresponding to the landslide and mudslide disaster risk based on the data point positions corresponding to each of the target terrain elevation data, the method also includes: obtaining manual inspection records for the risk area range; the manual inspection records define whether there is a landslide and mudslide disaster risk in the area identified during the manual inspection process; when the manual inspection records indicate that there is a landslide and mudslide disaster risk in the risk area range, determining the risk warning level corresponding to the risk warning operation according to the attribute information of the risk area range; the attribute information includes slope and / or area.

[0008] According to a landslide and mudslide disaster warning method provided by the present invention, the risk warning level corresponding to the risk warning operation is determined according to the attribute information of the risk area range, including: determining the regional risk confidence of the risk area range according to the confidence corresponding to each of the target terrain elevation data; determining the corresponding risk warning level according to the attribute information of the risk area range and the regional risk confidence.

[0009] According to a landslide and mudslide disaster warning method provided by the present invention, the risk warning operation is performed based on the risk area range, including: determining a target warning color matching the risk warning level from a plurality of preset warning colors; and marking the risk area range in the target area according to the target warning color.

[0010] According to a landslide and mudslide disaster warning method provided by the present invention, after obtaining the manual inspection record for the risk area, the method also includes: when the manual inspection record indicates that there is no landslide and mudslide disaster risk in the risk area, determining that there is no landslide and mudslide disaster risk in the risk area; determining the terrain elevation data sample with the corresponding label of no landslide and mudslide disaster risk according to the target terrain elevation data, so as to train and optimize the landslide and mudslide identification model.

[0011] According to a landslide and debris flow disaster warning method provided by the present invention, before inputting the terrain elevation dataset into a preset landslide and debris flow identification model, the method also includes: scaling the terrain elevation dataset according to the input data standard corresponding to the landslide and debris flow identification model.

[0012] According to a landslide and debris flow disaster early warning method provided by the present invention, after scaling the terrain elevation data set according to the input data standard corresponding to the landslide and debris flow identification model, the method further includes: detecting whether there is noise terrain elevation data corresponding to a blank area in the scaled terrain elevation data set; for each terrain elevation data in the terrain elevation data set, respectively configuring corresponding elevation validity flag information; the elevation validity flag information corresponding to the noise terrain elevation data is invalid; wherein, the terrain elevation data set is input into a preset landslide and debris flow identification model. The method comprises: inputting the terrain elevation data set and the corresponding elevation validity flag information into a preset landslide and debris flow identification model, so that the landslide and debris flow identification model processes each of the terrain elevation data based on a first channel, and processes each of the elevation validity flag information based on a second channel, so that the landslide and debris flow identification model identifies whether there is at least one target terrain elevation data corresponding to the landslide and debris flow disaster risk in the terrain elevation data set.

[0013] The present invention also provides a landslide and debris flow disaster warning device, which includes: an elevation data acquisition unit, which is used to acquire a terrain elevation data set corresponding to a target area to be identified; each terrain elevation data in the terrain elevation data set contains a corresponding data point position located in the target area; a model prediction unit, which is used to input the terrain elevation data set into a preset landslide and debris flow identification model to determine whether there is at least one target terrain elevation data corresponding to a landslide and debris flow disaster risk in the terrain elevation data set; each training sample in the sample data set of the landslide and debris flow identification model contains a terrain elevation data sample and a corresponding label, and the label is a landslide and debris flow disaster risk or a landslide and debris flow disaster risk; an early warning unit, which is used to perform an early warning operation based on each of the target terrain elevation data when it is determined that the at least one target terrain elevation data exists.

[0014] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the landslide and mud-rock flow disaster early warning method as described above is implemented.

[0015] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the landslide and mud-rock flow disaster warning method as described in any one of the above is implemented.

[0016] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the landslide and mud-rock flow disaster early warning method as described above is implemented.

[0017] The present invention provides a landslide and mud-rock flow disaster early warning method, device, electronic device and storage medium. In the method, a terrain elevation data set of a target area is obtained, and a landslide and mud-rock flow identification model is used to identify whether there is terrain data corresponding to the landslide and mud-rock flow disaster risk. The terrain characteristics of the landslide and mud-rock flow are mined by a machine learning method, thereby improving the recognition accuracy of the landslide and mud-rock flow disaster risk. Then, the terrain elevation data corresponding to the landslide and mud-rock flow disaster risk is used for early warning, which is also beneficial for operation and maintenance personnel to take facility protection measures for the risk area in a targeted manner in advance, thereby reducing property and personal safety losses caused by landslides and mud-rock flows. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 A flowchart showing an example of a landslide and debris flow disaster early warning method according to an embodiment of the present invention;

[0020] Figure 2 A flowchart showing another example of a landslide and debris flow disaster early warning method according to an embodiment of the present invention;

[0021] Figure 3 A flowchart showing another example of a landslide and debris flow disaster early warning method according to an embodiment of the present invention;

[0022] Figure 4 A schematic diagram showing an example of digital terrain point cloud data of a target area according to an embodiment of the present invention;

[0023] Figure 5 A schematic diagram showing the marked point cloud data corresponding to the landslide and debris flow disaster area according to an embodiment of the present invention;

[0024] Figure 6 A schematic diagram showing an example of a risk area range output by a landslide and debris flow identification model according to an embodiment of the present invention;

[0025] Figure 7 A schematic diagram showing an example of a disaster warning image for a risk area range of a target area according to an embodiment of the present invention;

[0026] Figure 8 A structural block diagram showing an example of a landslide and debris flow disaster warning device according to an embodiment of the present invention;

[0027] Fig. 9 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings 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.

[0029] It should be noted that the identification and early warning of landslide and debris flow disasters is a basic guarantee for the safe operation of field industrial facilities. At present, the commonly used identification method is mainly to collect the changes in the ground height of the monitoring area and predict it through mathematical models. Affected by the setting cost of the monitoring device and the error of data collection, the cost, accuracy and efficiency of the existing methods have room for further improvement.

[0030] Figure 1 A flow chart showing an example of a landslide and debris flow disaster early warning method according to an embodiment of the present invention.

[0031] Regarding the execution subject of the method of the embodiment of the present invention, it can be any controller with computing or processing capabilities to achieve the goal of accurately predicting and warning landslide and debris flow disasters. In some examples, it can be integrated and configured in an electronic device through software, hardware, or a combination of software and hardware, and can be a smart phone, smart watch, desktop computer, laptop, virtual reality terminal, augmented reality terminal, wireless terminal, laptop portable computer, edge computing node, cloud server, etc., which should not be limited here.

[0032] like Figure 1 As shown, in step S110, a terrain elevation data set corresponding to the target area to be identified is obtained.

[0033] Here, each terrain elevation data in the terrain elevation data set includes corresponding data point locations in the target area. Exemplarily, the terrain elevation data set can be determined by a DEM (Digital Elevation Model) file of the target area, and can be constructed by drone aerial photography.

[0034] In step S120, the terrain elevation data set is input into a preset landslide and debris flow identification model to determine whether there is at least one target terrain elevation data corresponding to the landslide and debris flow disaster risk in the terrain elevation data set.

[0035] Here, each training sample in the sample data set of the landslide and debris flow identification model includes a terrain elevation data sample and a corresponding label, and the label is that there is a landslide and debris flow disaster risk or there is no landslide and debris flow disaster risk.

[0036] It should be noted that the landslide and debris flow identification model can be a feature recognition network constructed using machine learning methods, and its structure is complex enough to support the mining of potential landslide and debris flow terrain features, without any specific restrictions on the specific structure of the neural network.

[0037] Specifically, the landslide and debris flow identification model mines and analyzes the terrain elevation data set according to the trained landslide and debris flow characteristics to identify whether there is a landslide and debris flow disaster risk at the location of each data point, and then completes the summary of the target terrain elevation data.

[0038] In step S130 , when it is determined that at least one target terrain elevation data exists, an early warning operation is performed based on each target terrain elevation data.

[0039] For example, when there is no target terrain elevation data, no early warning is required, but when there is target terrain elevation data, the target terrain elevation data is sent to the operation and maintenance personnel so that the operation and maintenance personnel can carry out risk prevention work in the target area as soon as possible.

[0040] Figure 2 A flowchart showing another example of a landslide and mud-rock flow disaster early warning method according to an embodiment of the present invention is shown.

[0041] like Figure 2 As shown, in step S210, a terrain elevation data set corresponding to the target area to be identified is obtained.

[0042] Here, the terrain elevation dataset can be data that can reflect the change of regional ground elevation, including but not limited to regional point cloud data, DEM file data, and elevation image data, etc. Exemplarily, by using convenient terrain acquisition equipment such as drone laser radar, the terrain elevation dataset corresponding to the target area to be identified is periodically collected.

[0043] In step S220, the terrain elevation data set is input into a preset landslide and debris flow identification model to determine whether there is at least one target terrain elevation data corresponding to the landslide and debris flow disaster risk in the terrain elevation data set.

[0044] In some embodiments, the terrain elevation dataset is scaled according to the input data standard corresponding to the landslide and debris flow identification model, thereby ensuring that the terrain elevation dataset meets the input data specification of the landslide and debris flow identification model through preprocessing operations. Exemplarily, the input data standard of the landslide and debris flow identification model is 416*416, and the terrain elevation dataset is scaled so that the long side of the area is mapped to 416 data points, and the short side is scaled at the same ratio, thereby meeting the above input data standard.

[0045] It should be noted that after the point cloud data is scaled, some noise data may be generated, such as blank areas caused by point cloud scaling. In some embodiments, it is detected whether there is noise terrain elevation data corresponding to the blank area in the scaled terrain elevation data set. Then, for each terrain elevation data in the terrain elevation data set, the corresponding elevation validity flag information is configured respectively, and the elevation validity flag information corresponding to the noise terrain elevation data is invalid, while the elevation validity flag information corresponding to other terrain elevation data is valid. Here, the elevation validity flag defines whether the elevation data at the data point location is valid, and the landslide and debris flow identification model is a multi-channel model.

[0046] Specifically, the terrain elevation data set and the corresponding elevation validation flag information are input into the preset landslide and debris flow identification model, so that the landslide and debris flow identification model processes the terrain elevation data based on the first channel, and processes the elevation validation flag information based on the second channel, so that the landslide and debris flow identification model identifies whether there is at least one target terrain elevation data corresponding to the landslide and debris flow disaster risk in the terrain elevation data set. Thus, the terrain elevation data is classified into valid data and invalid data, the meaningless data points generated by scaling are filtered out, the meaningful terrain elevation data are fully considered, and the prediction accuracy and processing efficiency of the landslide and debris flow identification model are improved.

[0047] In step S230, when it is determined that there is at least one target terrain elevation data, the risk area range corresponding to the landslide and mud-rock flow disaster risk is determined based on the data point positions corresponding to each target terrain elevation data.

[0048] Exemplarily, by connecting the data point positions in the elevation data corresponding to the landslide and debris flow risks, the risk area range corresponding to the landslide and debris flow disaster risk is drawn.

[0049] In step S240, a manual inspection record for the risk area is obtained, and the manual inspection record defines whether there is a risk of landslide and mud-rock flow disaster in the area identified during the manual inspection process.

[0050] In some embodiments, in order to ensure the reliability of risk identification results, after the risk area range is drawn up by machine learning through the landslide and debris flow identification model, the risk area range is manually inspected to manually identify whether there is a real risk of landslide and debris flow disasters.

[0051] In step S250, it is detected whether the manual inspection record indicates that there is a risk of landslide and mud-rock flow disaster in the risk area.

[0052] In step S261, when the manual inspection record indicates that there is a risk of landslide and debris flow disaster in the risk area, the risk warning level corresponding to the risk warning operation is determined according to the attribute information of the risk area, and the attribute information includes slope and / or area.

[0053] For example, when the area of ​​the risk zone is larger or the slope is steeper, the corresponding risk warning level is also higher.

[0054] In some implementations, the regional risk confidence of the risk area range is determined according to the confidence corresponding to each target terrain elevation data, and then the corresponding risk warning level is determined according to the attribute information of the risk area range and the regional risk confidence.

[0055] In some implementations, the average or median corresponding to the confidence of each target terrain elevation data is used as the regional risk confidence, and then the corresponding risk warning level is determined using the attribute information of the risk area range and the regional risk confidence. For example, when the regional risk confidence is low, the risk level can be directly determined as low risk, and when the regional risk confidence is high, the risk level is determined to be low risk or high risk based on the attribute information of the risk area range.

[0056] Furthermore, from multiple preset warning colors, a target warning color that matches the risk warning level is determined, and then the risk area range is marked in the target area according to the target warning color. For example, brown or red is used to mark the low-risk area range and the high-risk area range respectively, so as to intuitively reflect the risk area range and the corresponding risk level in the target area to the operation and maintenance personnel.

[0057] In step S263, when the manual inspection record indicates that there is no landslide and debris flow disaster risk in the risk area, it is determined that there is no landslide and debris flow disaster risk in the risk area.

[0058] In step S270, the terrain elevation data samples corresponding to the labels of no landslide and debris flow disaster risk are determined according to the target terrain elevation data to train and optimize the landslide and debris flow identification model. Thus, the training sample set of the landslide and debris flow identification model is improved through manual inspection records, so that the landslide and debris flow identification model is continuously iterated and optimized during the application of the model.

[0059] In the embodiment of the present invention, a landslide and debris flow disaster identification and early warning method based on machine learning is proposed. The terrain characteristics of landslide and debris flow are learned by the method based on machine learning to realize the identification of naturally occurring landslide and debris flow. In addition, by correcting and comprehensively evaluating the identification results, the risk early warning of the monitoring area is realized. At the same time, the evaluation results are corrected by introducing the results of manual inspections, which has the advantages of fast execution speed and high accuracy.

[0060] Through the embodiments of the present invention, the monitoring efficiency and accuracy of field landslide and mud-rock flow disasters are improved, the risk control capability of landslide and mud-rock flow disasters is improved, the problems of high cost and low accuracy of existing methods are solved, and the scheduling planning of risk elimination priorities is realized through multi-level risk warning guidance personnel.

[0061] Figure 3 A flow chart showing another example of a landslide and mud-rock flow disaster early warning method according to an embodiment of the present invention.

[0062] like Figure 3 As shown, the landslide and debris flow disaster early warning method includes a model building stage 310 , a risk identification stage 320 and a risk early warning stage 330 .

[0063] In the model building stage 310, training data is collected and marked. In order to train the landslide and debris flow disaster identification model, a large amount of training data is needed for the model. The terrain data of the detection area is obtained by means of laser radar mounted on unmanned aerial vehicles, and a digital terrain elevation data set reflecting the surface elevation information of the area is formed through processing. Then, the digital terrain elevation data set is processed by manual screening and marking, and the areas where landslide and debris flow disasters exist are marked to form a digital terrain disaster marking data set. Here, each digital terrain elevation data and its corresponding marking data correspond one-to-one to form a set of training data, and a large number of training data constitute a model training data set.

[0064] In some embodiments, a digital terrain elevation data set is obtained through a landslide and debris flow data simulation method. For example, a large number of digital terrain point cloud files are downloaded through the Internet, and a total of a groups of simulation parameters are set for data simulation according to the characteristics of the terrain files. A group of digital terrain elevation data is obtained through the landslide and debris flow simulation method, and a corresponding digital terrain disaster marker data set is synchronously obtained during the simulation process, which is used to represent the point cloud data of the landslide and debris flow area.

[0065] Figure 4 A schematic diagram showing an example of digital terrain point cloud data of a target area according to an embodiment of the present invention. Figure 5 A schematic diagram of marked point cloud data corresponding to a landslide and debris flow disaster area according to an embodiment of the present invention is shown.

[0066] Furthermore, after obtaining the model training data set, in order to increase the number of training samples and improve the model training effect, this embodiment performs data enhancement on the training data set, rotates the data elevation data and its corresponding disaster marker data by 90°, 180°, and 270°, respectively, and expands the number of training samples to four times the original number of samples, that is, a total of 4a groups of training sample data.

[0067] Then, the landslide and debris flow recognition model is trained based on the model training data set. Here, a neural network can be built according to the network structure of YOLOv3, and the input specification of the network is set to 416*416*2, where channel 1 is digital terrain elevation data processed to a size of 416*416, each data is a regional interpolation elevation value, and channel 2 is also 416*416 scale data, which is the effective flag corresponding to the elevation, marking whether the data here is a meaningful value, 1 is meaningful, and 0 is meaningless. In some examples, the output of the network structure is divided into three scales according to the size of the detected target, namely 13*13*18, 26*26*18 and 52*52*18. From the data available, randomly shuffle and take out 80% as a training set for network training, take out 10% as a validation set, and the remaining 10% as a test set. Specifically, the neural network constructed for this example can be trained for a total of b rounds using the training data set. The network model finally obtained has the ability to identify landslide and mudslide disasters and can mark the areas and confidence levels where disasters occur. This completes the construction of the landslide and mudslide identification model. The neural network is trained using a large amount of data to learn the terrain features of the landslide and mudslide marked areas, so that the network has the ability to identify landslide and mudslide disasters through digital elevation data.

[0068] In the risk identification stage 320, a copy of the regional terrain elevation data of a certain area to be identified is selected and input into the landslide and debris flow identification model to obtain the model's identification result for this area. In order to enable the regional terrain elevation data to be input into the network, it needs to be processed into an input specification of 416*416*2, where channel 1 is the interpolated elevation value of the regional terrain elevation data, with a specification of 416*416, and channel 2 stores the validity of the elevation data of this area. If a certain elevation data is valid, it is 1, and if it is invalid, it is saved as 0, and the data specification is also 416*416. Exemplarily, when the selected regional data is not a square, in order to ensure that the shape of the data does not change, it is necessary to stretch the length and width directions in equal proportions so that the long side of the area is mapped to 416 data points, and the short side is scaled at the same ratio.

[0069] For some blank areas caused by scaling, the elevation data uses 0 as a placeholder, and the validity data is set to 0, that is, the data here is invalid; the above data is input into the recognition model, and the output is obtained after calculation; the output of the network marks the possible recognition type L, target center x, y, target size w, h, and confidence c. Then, the target center, size, and confidence of the target recognition type of the landslide and debris flow area are extracted as the result of disaster risk identification.

[0070] Specifically, the scaling process may adopt the following operation, wherein the input data is the digital elevation data to be scaled, and the output data is the data that meets the network input condition, such as the data that meets the data size of 416*416:

[0071] gap=416-1

[0072]

[0073] ori x =((w-1)-step*gap) / 2

[0074] ori y =((h-1)-step*gap) / 2

[0075] height xy =ori_hright[ori x +step*x][ori y +step*y]

[0076] Among them, gap is the number of gaps in each dimension of the output data, that is, 415 gaps; step represents the coordinate span in the input data when the output data is separated by 1 unit; ori x ,ori y They are the coordinate positions of the origin position in the output data corresponding to the input data; height xy is the data value with coordinates x and y in the output data; ori_hright[x][y] is the data value with coordinates x and y in the input data. If the coordinates are illegal, they are set to 0. If the coordinates are not integers, the corresponding area is interpolated. For the data of channel 2, scaling is achieved in the same way.

[0077] Furthermore, the inspection data set obtained through the manual inspection process is monitored to calibrate the output results of the disaster identification model. When the manual inspection results are different from the model output results, they will be corrected according to the manual inspection results. The corrected identification results will be marked according to their risk characteristics to obtain the corrected landslide and debris flow risk prediction results. If there are no manual inspection results in the risk area, no correction will be performed.

[0078] In the risk warning stage 330, when it is identified that there is no manual inspection data for the corresponding area within the risk area range, there is no need to correct the identification result, and the warning area is directly marked according to the output result of the model.

[0079] In some embodiments, the confidence of the model prediction results is comprehensively considered to complete the prediction of the risk area. For example, when the confidence is less than 50%, it is determined to be low risk, and when the confidence is higher than 50%, the risk level of the risk area range is determined with reference to Table 1:

[0080] Table 1 Risk level identification table

[0081]

[0082] Figure 6 FIG. 1 is a schematic diagram showing an example of a risk area range output by a landslide and debris flow identification model according to an embodiment of the present invention. Figure 6 In the example in , the corresponding risk identification confidence is 72%. Since the landslide area is medium in size and has a gentle slope, it is marked as a medium-risk area.

[0083] Figure 7 A schematic diagram showing an example of a disaster warning image for a risk area range of a target area according to an embodiment of the present invention.

[0084] like Figure 7 As shown, the landslide area in the target area is marked as a medium-risk area. According to the need, the medium-risk area in the identified image is painted in a color different from other surrounding areas. For example, other risk-free areas are marked in green, while the medium-risk area is painted in yellow. Thus, a disaster warning image is obtained by coloring the warning area of ​​the identified image.

[0085] Through the embodiments of the present invention, landslide and mud-rock flow disasters are identified based on machine learning models. By using computer platforms and machine learning techniques, a large amount of training data is learned to build a disaster identification model. The trained model can identify the landslide and mud-rock flow areas existing therein according to the digital terrain data. Then, the model identification results are corrected with reference to the manual inspection results, and the corrected results are divided into different risk levels according to their characteristics to achieve guidance on processing priorities. In addition, the identification results are displayed in the form of disaster warning images to improve the efficiency of information transmission. As a result, the identification and warning of landslide and mud-rock flow disasters occurring in the wild area can be realized more efficiently and accurately.

[0086] Figure 8 A structural block diagram of an example of a landslide and mud-rock flow disaster warning device according to an embodiment of the present invention is shown.

[0087] like Figure 8 As shown, the landslide and debris flow disaster warning device 800 includes an elevation data acquisition unit 810 , a model prediction unit 820 and a warning unit 830 .

[0088] The elevation data acquisition unit 810 is used to acquire a terrain elevation data set corresponding to the target area to be identified; each terrain elevation data in the terrain elevation data set includes a corresponding data point position located in the target area.

[0089] The model prediction unit 820 is used to input the terrain elevation data set into a preset landslide and debris flow identification model to determine whether there is at least one target terrain elevation data corresponding to a landslide and debris flow disaster risk in the terrain elevation data set; each training sample in the sample data set of the landslide and debris flow identification model contains a terrain elevation data sample and a corresponding label, and the label is that there is a landslide and debris flow disaster risk or there is no landslide and debris flow disaster risk.

[0090] The early warning unit 830 is configured to perform an early warning operation based on each of the target terrain elevation data when it is determined that the at least one target terrain elevation data exists.

[0091] Fig. 9 An example of a physical structure diagram of an electronic device is shown in FIG. Fig. 9 As shown, the electronic device may include: a processor (processor) 910, a communication interface (Communications Interface) 920, a memory (memory) 930 and a communication bus 940, wherein the processor 910, the communication interface 920, and the memory 930 communicate with each other through the communication bus 940. The processor 910 can call the logic instructions in the memory 930 to execute a landslide and debris flow disaster warning method, which includes: obtaining a terrain elevation data set corresponding to a target area to be identified; each terrain elevation data in the terrain elevation data set contains a corresponding data point position located in the target area; the terrain elevation data set is input into a preset landslide and debris flow identification model to determine whether there is at least one target terrain elevation data corresponding to a landslide and debris flow disaster risk in the terrain elevation data set; each training sample in the sample data set of the landslide and debris flow identification model contains a terrain elevation data sample and a corresponding label, and the label is a landslide and debris flow disaster risk or a landslide and debris flow disaster risk; when it is determined that the at least one target terrain elevation data exists, a warning operation is performed based on each of the target terrain elevation data.

[0092] In addition, the logic instructions in the above-mentioned memory 930 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0093] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the landslide and mudslide disaster warning method provided by the above-mentioned methods, and the method includes: obtaining a terrain elevation data set corresponding to a target area to be identified; each terrain elevation data in the terrain elevation data set contains a corresponding data point position located in the target area; the terrain elevation data set is input into a preset landslide and mudslide identification model to determine whether there is at least one target terrain elevation data corresponding to a landslide and mudslide disaster risk in the terrain elevation data set; each training sample in the sample data set of the landslide and mudslide identification model contains a terrain elevation data sample and a corresponding label, and the label is a landslide and mudslide disaster risk or a landslide and mudslide disaster risk; when it is determined that there is at least one target terrain elevation data, a warning operation is performed based on each of the target terrain elevation data.

[0094] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the landslide and mudslide disaster warning method provided by the above-mentioned methods, the method comprising: obtaining a terrain elevation data set corresponding to a target area to be identified; each terrain elevation data in the terrain elevation data set contains a corresponding data point position located in the target area; the terrain elevation data set is input into a preset landslide and mudslide identification model to determine whether there is at least one target terrain elevation data corresponding to a landslide and mudslide disaster risk in the terrain elevation data set; each training sample in the sample data set of the landslide and mudslide identification model contains a terrain elevation data sample and a corresponding label, the label being a landslide and mudslide disaster risk or a landslide and mudslide disaster risk; when it is determined that the at least one target terrain elevation data exists, a warning operation is performed based on each of the target terrain elevation data.

[0095] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0096] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0097] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A landslide and debris flow disaster early warning method, characterized in that: The method comprises: Acquire a terrain elevation data set corresponding to the target area to be identified; each terrain elevation data in the terrain elevation data set includes a corresponding data point position located in the target area; The terrain elevation data set is input into a preset landslide and debris flow identification model to determine whether there is at least one target terrain elevation data corresponding to a landslide and debris flow disaster risk in the terrain elevation data set; each training sample in the sample data set of the landslide and debris flow identification model comprises a terrain elevation data sample and a corresponding label, wherein the label is a landslide and debris flow disaster risk or a non-landslide and debris flow disaster risk; When it is determined that the at least one target terrain elevation data exists, an early warning operation is performed based on each of the target terrain elevation data.

2. The landslide and debris flow disaster early warning method according to claim 1, characterized in that: The performing of the early warning operation based on each of the target terrain elevation data comprises: Determine the risk area range corresponding to the landslide and mud-rock flow disaster risk based on the data point positions corresponding to each of the target terrain elevation data; Based on the risk area range, a risk warning operation is performed.

3. The landslide and debris flow disaster early warning method according to claim 2 is characterized in that: After determining the risk area range corresponding to the landslide and mud-rock flow disaster risk based on the data point positions corresponding to each of the target terrain elevation data, the method further includes: Obtaining manual inspection records for the risk area; the manual inspection records define whether there is a risk of landslide and mud-rock flow disaster in the area identified during the manual inspection process; When the manual inspection record indicates that there is a risk of landslide and mud-rock flow disaster in the risk area, the risk warning level corresponding to the risk warning operation is determined according to the attribute information of the risk area; the attribute information includes slope and / or area.

4. The landslide and debris flow disaster early warning method according to claim 3 is characterized in that: The determining, according to the attribute information of the risk area range, the risk warning level corresponding to the risk warning operation includes: Determining the regional risk confidence level of the risk area range according to the confidence level corresponding to each target terrain elevation data; The corresponding risk warning level is determined according to the attribute information of the risk area range and the regional risk confidence.

5. The landslide and debris flow disaster early warning method according to claim 3 is characterized in that: The performing of risk warning operations based on the risk area range includes: Determining a target warning color that matches the risk warning level from a plurality of preset warning colors; The risk area range is marked in the target area according to the target warning color.

6. The landslide and mud-rock flow disaster early warning method according to claim 3 is characterized in that: After obtaining the manual inspection record for the risk area, the method further includes: When the manual inspection record indicates that there is no landslide and debris flow disaster risk in the risk area, it is determined that there is no landslide and debris flow disaster risk in the risk area; According to the target terrain elevation data, terrain elevation data samples corresponding to labels of no landslide and debris flow disaster risk are determined to train and optimize the landslide and debris flow identification model.

7. The landslide and debris flow disaster early warning method according to claim 1, characterized in that: Before inputting the terrain elevation data set into a preset landslide and debris flow identification model, the method further includes: The terrain elevation data set is scaled according to the input data standard corresponding to the landslide and debris flow identification model.

8. The landslide and debris flow disaster early warning method according to claim 7, characterized in that: After scaling the terrain elevation data set according to the input data standard corresponding to the landslide and debris flow identification model, the method further includes: Detect whether there is noise terrain elevation data corresponding to the blank area in the scaled terrain elevation dataset; For each terrain elevation data in the terrain elevation data set, corresponding elevation validity flag information is configured respectively; the elevation validity flag information corresponding to the noise point terrain elevation data is invalid; The step of inputting the terrain elevation data set into a preset landslide and debris flow identification model to determine whether there is at least one target terrain elevation data corresponding to a landslide and debris flow disaster risk in the terrain elevation data set includes: The terrain elevation data set and the corresponding elevation validity flag information are input into a preset landslide and debris flow identification model, so that the landslide and debris flow identification model processes each terrain elevation data based on the first channel and processes each elevation validity flag information based on the second channel, so that the landslide and debris flow identification model can identify whether there is at least one target terrain elevation data corresponding to the landslide and debris flow disaster risk in the terrain elevation data set.

9. A landslide and debris flow disaster early warning device, characterized in that: The device comprises: An elevation data acquisition unit, used to acquire a terrain elevation data set corresponding to a target area to be identified; each terrain elevation data in the terrain elevation data set includes a corresponding data point position located in the target area; A model prediction unit is used to input the terrain elevation data set into a preset landslide and debris flow identification model to determine whether there is at least one target terrain elevation data corresponding to a landslide and debris flow disaster risk in the terrain elevation data set; each training sample in the sample data set of the landslide and debris flow identification model includes a terrain elevation data sample and a corresponding label, and the label is a landslide and debris flow disaster risk or a landslide and debris flow disaster risk. The early warning unit is used to perform an early warning operation based on each of the target terrain elevation data when it is determined that the at least one target terrain elevation data exists.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the landslide and mud-rock flow disaster early warning method as claimed in any one of claims 1 to 8 is implemented.

Images