Remote sensing monitoring method and system for potential safety hazards along railway based on geographic information

Abstract

The invention provides a remote sensing monitoring method and system for potential safety hazards along a railway based on geographic information, and the method comprises the steps: obtaining remote sensing image data related to each potential safety hazard target at different time and / or different data sources for potential safety hazard targets in a predetermined railway surrounding environment; utilizing a pre-trained generative adversarial network used for identifying railway surrounding environment potential safety hazards to extract features from the remote sensing image data and generate potential safety hazard identification results; and integrating potential safety hazard identification results obtained by the remote sensing image data related to each potential safety hazard target at different time, and realizing the potential safety hazard monitoring of the railway surrounding environment based on the multi-temporal remote sensing image data in combination with geocoding and registration results of geographic information contained in the labels of the remote sensing image data. According to the invention, potential safety hazard monitoring of the railway surrounding environment based on the multi-temporal remote sensing image data can be realized.

Description

Technical Field

[0001] The present invention relates to the field of railway safety information technology, and in particular to a method and system for remote sensing monitoring of potential safety hazards along railways based on geographic information. Background Art

[0002] At present, satellite remote sensing images have a large coverage area and low acquisition cost. With the development and application of my country's high-resolution satellite data platform, the spatial resolution of remote sensing data has been significantly improved, and multi-spectral remote sensing image data with high spatial resolution of meters and sub-meters can be obtained. Therefore, it is possible to use satellite remote sensing data to monitor environmental safety hazards along high-speed railways.

[0003] However, in the actual application of satellite remote sensing technology, there is a lack of accuracy in identifying environmental safety hazards along the railway. At the same time, there is a lack of deeper integration of satellite remote sensing imaging technology and professional information in the railway industry, and the monitoring and identification results lack the association of information on important infrastructure in the railway industry (such as important bridges, tunnels and mileage markers, etc.). Summary of the invention

[0004] In view of this, an embodiment of the present invention provides a method and system for remote sensing monitoring of safety hazards along a railway based on geographic information, so as to eliminate or improve one or more defects existing in the prior art.

[0005] One aspect of the present invention provides a method for remote sensing monitoring of safety hazards along a railway based on geographic information, the method comprising the following steps: for predetermined safety hazard targets in the surrounding environment of the railway, remote sensing image data at different times and / or different data sources related to each safety hazard target are obtained; wherein the remote sensing image data has undergone preprocessing including correction processing and standardization processing, and the label of each remote sensing image data contains geographic coding and registration results based on railway-specific geographic information; using a pre-trained generative adversarial network for identifying safety hazards in the surrounding environment of the railway to extract features from each remote sensing image data and generate safety hazard identification results; wherein the generative adversarial network is trained using a pre-constructed safety hazard element target hazard sample library as real samples, and the safety hazard element target hazard sample library contains remote sensing image data and corresponding target hazard information; the safety hazard identification results obtained from the remote sensing image data at different times related to each safety hazard target are integrated, combined with the geographic coding and registration results of the geographic information contained in the label of the remote sensing image data, to realize safety hazard monitoring of the surrounding environment of the railway based on multi-temporal remote sensing image data.

[0006] In some embodiments of the present invention, the preprocessing of the remote sensing image data also includes multiple processings such as data mosaicking, bad pixel repairing, color uniformity and image cropping; the correction processing includes radiation correction and orthorectification; and the standardization processing includes unifying feature quantities of different scales carried in the remote sensing image data to the same scale.

[0007] In some embodiments of the present invention, the method also includes the step of fusing remote sensing image data with geographic information data, calling the railway-specific geographic information of the railway geographic information system to annotate the remote sensing image data with labels containing geographic coding and alignment results, the step including: converting the vector data contained in the remote sensing image data of each safety hazard target into one or more of the WGS84 geographic coordinate system, the CGCS2000 geographic coordinate system, the Albers plane coordinate system and the UTM projection, calling the railway-specific geographic information of the railway geographic information system for annotation, and using the unique identifier of the geographic element as the annotation of the geographic coding and alignment results of each remote sensing image data.

[0008] In some embodiments of the present invention, the step of realizing safety hazard monitoring of the railway surrounding environment based on multi-temporal remote sensing image data by combining the geographic information contained in the label of the remote sensing image data with the geographic coding and registration results includes: determining the position of the safety hazard target corresponding to the remote sensing image data with safety hazards in the railway surrounding environment in WGS84 geographic coordinate system, CGCS2000 geographic coordinate system, Albers plane coordinate system or UTM projection based on the geographic coding and registration results of the geographic information contained in the label of the remote sensing image data, and realizing the visualization display of multi-temporal remote sensing images around the railway based on the statistics of multi-temporal remote sensing image data, and realizing the time series retrieval of spatial information of ground objects and the positioning and retrieval of environmental change information in the railway surrounding environment based on the statistics of multi-temporal remote sensing image data.

[0009] In some embodiments of the present invention, the method also includes the step of pre-constructing a safety hazard element target hazard sample library, which step includes: performing image cropping on the remote sensing image data corresponding to each safety hazard target; receiving target hazard information annotated by personnel on different cropped remote sensing image data; based on the remote sensing image data and target hazard information annotated by personnel, constructing a sample label migration network model; wherein the sample label migration network model is used to generate corresponding safety hazard identification results based on the input remote sensing image data.

[0010] In some embodiments of the present invention, the target hidden danger information includes a safety hidden danger category and a safety hidden danger level, and the safety hidden danger identification result includes a safety hidden danger category and a safety hidden danger level.

[0011] In some embodiments of the present invention, the method also includes a pre-training step of generating an adversarial network, which step includes: using a pre-constructed safety hazard factor target hazard sample library as real samples, using a generator to generate false samples, and training in an iterative process to improve the discriminator's ability to distinguish between the real samples and the false samples, thereby achieving pre-training of the generative adversarial network.

[0012] Corresponding to the above method, the present invention also provides a remote sensing monitoring system for safety hazards along railways based on geographic information, including a processor, a memory, and a computer program / instructions stored in the memory, wherein the processor is used to execute the computer program / instructions, and when the computer program / instructions are executed, the device / system implements the steps of the method described in any one of the above embodiments.

[0013] Corresponding to the above method, the present invention also provides a computer-readable storage medium on which a computer program / instruction is stored. When the computer program / instruction is executed by a processor, the steps of the method described in any one of the above embodiments are implemented.

[0014] Corresponding to the above method, the present invention also provides a computer program product, including a computer program / instruction, which implements the steps of the method described in any one of the above embodiments when executed by a processor.

[0015] The method and system for remote sensing monitoring of safety hazards along railways based on geographic information proposed in the present invention can identify safety hazards based on multi-temporal remote sensing image data acquired for different safety hazard targets using a pre-trained generative adversarial network, and can monitor safety hazards in the surrounding environment of the railway based on multi-temporal remote sensing image data by integrating the safety hazard identification results identified in different phases.

[0016] Additional advantages, purposes, and features of the present invention will be described in part in the following description, and will become apparent to those skilled in the art after studying the following, or may be learned from the practice of the present invention. The purposes and other advantages of the present invention may be achieved and obtained by the structures specifically indicated in the specification and the accompanying drawings.

[0017] Those skilled in the art will appreciate that the objectives and advantages that can be achieved with the present invention are not limited to the above specific description, and the above and other objectives that can be achieved by the present invention will be more clearly understood from the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the present invention, constitute a part of the present application, and do not constitute a limitation of the present invention. In the drawings:

[0019] Figure 1 The present invention is a flowchart of a method for remote sensing monitoring of safety hazards along railways based on geographic information in one embodiment of the present invention.

[0020] Figure 2 FIG. 1 is a schematic diagram of the structure of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0022] It should also be noted that, in order to avoid obscuring the present invention due to unnecessary details, only structures and / or processing steps closely related to the solutions according to the present invention are shown in the accompanying drawings, while other details that are not closely related to the present invention are omitted.

[0023] It should be emphasized that the term “include / comprises” when used herein refers to the presence of features, elements, steps or components, but does not exclude the presence or addition of one or more other features, elements, steps or components.

[0024] It should also be noted that, unless otherwise specified, the term “connection” herein may refer not only to a direct connection but also to an indirect connection involving an intermediate.

[0025] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the accompanying drawings, the same reference numerals represent the same or similar components, or the same or similar steps.

[0026] In order to overcome the problems existing in the prior art, the present invention proposes a remote sensing monitoring method and system for safety hazards along railways based on geographic information. In order to accurately identify safety hazards in the surrounding environment along railways that affect railway driving safety, the typical target feature extraction method of safety hazards in the surrounding environment along railways is improved, and the accuracy of the extracted features is improved, thereby achieving more accurate monitoring of safety hazards along railways based on satellite remote sensing image recognition.

[0027] Figure 1 This is a flow chart of a method for remote sensing monitoring of safety hazards along railways based on geographic information in an embodiment of the present invention. The method comprises the following steps:

[0028] Step S110: For predetermined safety hazard targets in the surrounding environment of the railway, remote sensing image data related to each safety hazard target at different times and / or different data sources are obtained; wherein the remote sensing image data has undergone preprocessing including correction processing and standardization processing, and the label of each remote sensing image data includes geographic coding and alignment results based on railway-specific geographic information.

[0029] Among them, the types of potential safety hazard targets include but are not limited to dust screens and color steel tiles.

[0030] Among them, remote sensing image data at different times and / or different data sources belong to multi-temporal data. Multi-temporal usually refers to the characteristics of a set of remote sensing images in a time series. Broadly speaking, any set of images, maps or geographic data of the same region acquired at different times can be regarded as "multi-temporal" data. Multi-temporal remote sensing data refers to image data with multiple time resolutions collected through repeated observations of the ground at a specific period, which can meet the needs of monitoring and analyzing dynamic changes.

[0031] Step S120: Utilize a pre-trained generative adversarial network for identifying safety hazards in the surrounding environment of railways to extract features from each remote sensing image data and generate safety hazard identification results; wherein the generative adversarial network is trained using a pre-constructed safety hazard element target hazard sample library as real samples, and the safety hazard element target hazard sample library contains remote sensing image data and corresponding target hazard information.

[0032] Step S130: Integrate the safety hazard identification results obtained from the remote sensing image data at different times related to each safety hazard target, and combine the geocoding and registration results of the geographic information contained in the labels of the remote sensing image data to realize safety hazard monitoring of the railway surrounding environment based on multi-temporal remote sensing image data.

[0033] By adopting the embodiment of the invention, the method and system for remote sensing monitoring of safety hazards along railways based on geographic information proposed in the present invention can identify safety hazards based on multi-temporal remote sensing image data acquired for different safety hazard targets using a pre-trained generative adversarial network, and can realize safety hazard monitoring of the railway surrounding environment based on multi-temporal remote sensing image data by integrating the safety hazard identification results identified in different phases.

[0034] In some embodiments of the present invention, the preprocessing of remote sensing image data further includes multiple processes such as data mosaic processing, bad pixel repair processing, color uniformity processing, and image cropping processing. Among them, the correction processing includes radiation correction and orthorectification processing, and the standardization processing includes unifying the feature quantities of different scales carried in the remote sensing image data to the same scale.

[0035] In the specific implementation process, the implementation method of step S110 can be: taking the dust screen and color steel tiles of the railway surrounding environment as the main target, select multi-temporal high-resolution remote sensing images and data. Preprocess the acquired remote sensing images of the high-speed railway surrounding environment, where the preprocessing steps include: data reading, radiation correction and orthorectification; mosaicking data, mosaicking a large number of large, high-precision orthorectified remote sensing data into one file according to the mosaic data set definition rules; repairing bad points of remote sensing data over a large area; performing uniform color processing on the uneven colors of adjacent remote sensing data that are connected or overlapped; and completing cropping and cutting 2.5km away from the center line of the railway for the areas covered along the railway.

[0036] In the specific implementation process, the standardized processing of multi-source remote sensing image data can be achieved based on the following methods. In order to solve the correlation problem between data of different scales and phases, the synergistic relationship between the image resolution, monitoring period, imaging band and hidden danger feature elements of different sensor platforms is extracted, and the feature quantities of different scales in multi-source data are unified to the same scale to increase the generalization ability of the data model.

[0037] The normalization calculation for the feature quantity is as follows:

[0038]

[0039] The minimum and maximum values ​​of each feature dimension are proportionally scaled to the range of [-1, +1], and each layer of data is standardized during preprocessing to achieve effective fusion of multi-source data.

[0040] Among them, data mosaic technology, especially in the field of remote sensing and image processing, usually refers to the technology of stitching multiple adjacent digital images into a large image, also known as digital mosaic. This technology is particularly important when dealing with areas beyond the coverage of a single image. It can provide a complete geographic information perspective, integrate multiple spatiotemporal data sources, and provide support for scientific research, resource management, and geographic information applications. The cropping step refers to first collecting remote sensing images of the area near the railway line, and then in the last step of preprocessing, cropping to obtain a cropped image with a preset distance from the center line of the railway.

[0041] In some embodiments of the present invention, the method also includes the step of fusing remote sensing image data with geographic information data, calling the railway-specific geographic information of the railway geographic information system to annotate the remote sensing image data with labels containing geographic coding and registration results, the step including: converting the vector data contained in the remote sensing image data of each safety hazard target into one or more of the WGS84 geographic coordinate system, the CGCS2000 geographic coordinate system, the Albers plane coordinate system and the UTM projection, calling the railway-specific geographic information of the railway geographic information system for annotation, and using the unique identifier of the geographic element as the annotation of the geographic coding and alignment results of each remote sensing image data.

[0042] In the specific implementation process, the vector data of surface elements such as dust-proof nets can be converted from the original coordinate system to the CGCS2000 (China Geodetic Coordinate System 2000) geographic coordinate system to match the positioning and analysis of the entire railway line. The dust-proof net vector data can also be converted to the WGS84 geographic coordinate system, and the GIS spatial statistical analysis tool can be used to extract the central longitude and latitude coordinates of each surface element (such as each dust-proof net target); or the vector data can be converted to the Albers plane coordinate system to calculate the area information of each surface element; or the vector data can be converted to the UTM projection to calculate the shortest distance of each surface element from the high-speed railway trunk line.

[0043] Accordingly, in some embodiments of the present invention, in combination with different geographic coordinate systems, the geocoding and registration results of the geographic information contained in the label of the remote sensing image data are combined in step S130 to implement the step of monitoring safety hazards in the surrounding environment of the railway based on multi-temporal remote sensing image data, including: based on the geocoding and registration results of the geographic information contained in the label of the remote sensing image data, determining the position of the safety hazard target corresponding to the remote sensing image data with safety hazards in the surrounding environment of the railway in the WGS84 geographic coordinate system, CGCS2000 geographic coordinate system, Albers plane coordinate system or UTM projection, and based on the statistics of the multi-temporal phase remote sensing image data, realizing the visualization display of multi-temporal remote sensing images around the railway, and based on the statistics of the multi-temporal phase remote sensing image data, realizing the time series retrieval of spatial information of ground objects in the surrounding environment of the railway and the positioning and retrieval of environmental change information.

[0044] By adopting the embodiment of the invention, unified intelligent management of time and space dimensions can be performed through geocoding and registration, thereby solving the spatial correlation problem between safety hazards and railway equipment and facilities.

[0045] In some embodiments of the present invention, the method also includes the step of pre-constructing a sample library of safety hazard elements and target hazards, which includes: (1) performing image cropping on remote sensing image data corresponding to each safety hazard target; (2) receiving target hazard information annotated by personnel on remote sensing image data after different cropping; (3) constructing a sample label transfer network model based on the remote sensing image data and target hazard information annotated by personnel; wherein the sample label transfer network model is used to generate corresponding safety hazard identification results based on the input remote sensing image data.

[0046] The target hidden danger information includes the safety hidden danger category and the safety hidden danger level, and the safety hidden danger identification result includes the safety hidden danger category and the safety hidden danger level.

[0047] By adopting this embodiment of the invention, unified intelligent management of time / space dimensions can be utilized through geocoding and registration, thereby solving the spatial correlation problem between safety hazards and railway equipment and facilities.

[0048] In some embodiments of the present invention, the method also includes a pre-training step of generating an adversarial network, which step includes: using a pre-constructed safety hazard factor target hazard sample library as real samples, using a generator to generate false samples, and training in an iterative process to improve the discriminator's ability to distinguish between the real samples and the false samples, thereby achieving pre-training of the generative adversarial network.

[0049] By adopting the embodiment of the invention, it is possible to use a generative adversarial network to distinguish between remote sensing image data with safety hazard elements and remote sensing image data without safety hazard elements. By using the proposed sample library annotation and amplification method based on the transfer learning algorithm, it is possible to solve the problem of small data sets of environmental hazards around railways, so that the model trained based on this method can cover various types of safety hazards along the railway.

[0050] In some embodiments of the present invention, a target hidden danger sample library of safety hidden danger elements is constructed, and all-element target hidden danger samples of high-resolution image data can be selected for fused multi-source remote sensing collaborative image data.

[0051] In some other embodiments of the present invention, remote sensing data of the railway periphery from data sources with different resolutions may be fused, and the data with different resolutions may be subsequently processed uniformly through standardization, thereby achieving fusion of multi-source remote sensing data.

[0052] For example, in order to improve the generalization ability of network segmentation targets and adapt to the long linear strip characteristics of railways, the target image can be cropped to 256×256 pixels, annotated by visual interpretation, and a round of annotation inspection can be performed. The sample library construction technology based on the transfer learning algorithm obtains remote sensing images containing safety hazard targets in different environments, time or scales, establishes a sample label migration network model, takes the existing safety hazard samples and heterogeneous remote sensing images of new scenes as the input of the network, and outputs new sample labels of heterogeneous remote sensing images after feature space conversion.

[0053] By adopting the embodiment of the invention, a safety hazard element target hazard sample library can be constructed based on multi-temporal remote sensing image data.

[0054] In some embodiments of the present invention, feature extraction is performed through a pre-trained GAN deep learning network.

[0055] For example, a GAN (Generative Adversarial Network, GAN) deep learning network can be constructed based on the deep feature extraction learning process of multi-temporal data. Combined with the changing factors of safety hazards, a GAN loss function suitable for the target scenario of safety hazards in the railway surrounding environment is designed. In the GAN loss function, D is a fixed discriminator and G is a training generator. When the generator maximizes the cross entropy loss V(D,G) of true and false images in the discriminator, the minimum cross entropy loss min max V(D,G) is calculated. Among them, the discriminator loss function (D_loss) and the generator loss function (G_loss)

[0056] min G max D V(D,G)=E xT [logD(x)]+E zF [log 1-D(G(z))]

[0057] D(x) represents the probability that x is a real target. If it is 1, it means that it is 100% a real target feature, and if the output is 0, it represents a non-feature target. The V(D,G) on the left side of the formula represents the difference between the generated image and the target feature, and adopts the cross entropy loss function of binary classification (true and false categories). It contains two parts: minG and maxD: max V(D,G) means that the fixed generator G trains the discriminator D, and the parameters of the discriminator D are updated by maximizing the cross entropy loss V(D,G). The training goal of D is to correctly distinguish between the real image D x and the generated image G(z). The stronger the identification ability of D, the larger D(x) should be, the larger the first term on the right, and the smaller D(G(x)) should be, and the second term on the right is larger. At this time, V(D,G) will become larger, so the formula for D is to find the maximum (maxD).

[0058] By adopting the embodiment of the invention, the constructed safety hazard factor target hazard sample library can be used to train and optimize the network. The deep learning network uses labeled data for repeated training to implement a good discriminant model for change detection, which can effectively reduce the dependence on the amount of data and improve the network robustness.

[0059] Furthermore, the sample library annotation and amplification method based on the transfer learning algorithm can be used to solve the problem of the small number of data sets of hidden dangers in the railway surrounding environment, so that the model trained based on this method can cover various types of safety hazards along the railway.

[0060] In some embodiments of the present invention, image coding and registration steps based on geographic information are included. It aims to integrate multi-source remote sensing images and geographic information data, and can call railway-specific geographic information of the railway geographic information system for geographic coding and registration. According to the identified safety hazard target parameters, through spatial feature extraction and image analysis, scanning, detection, and recognition are performed, and the output detection and recognition result data including relevant images, location information, feature information, etc. are calculated and output; the relevant ledgers, location information and feature information are associated with the existing important infrastructure equipment of the railway, and the dust net vector data is converted to the WGS84 geographic coordinate system, and the GIS spatial statistical analysis tool is used to extract the central longitude and latitude coordinates of each surface element (such as each dust net target); the vector data is converted to the Albers plane coordinate system, and the area information of each surface element is calculated; the vector data is converted to the UTM projection, and the shortest distance of each surface element from the high-speed rail trunk line is calculated.

[0061] Furthermore, the above information can be stored in the attribute list of the vector data and exported to Excel format together with the FID number to achieve further statistical analysis, field verification and data integration. Among them, the FID number is the FID in ArcGIS, which is a unique identification method for geographic information.

[0062] In another embodiment of the present invention, a multi-source information fusion processing method based on remote sensing data combined with GIS can be used to check external railway safety hazards and safety protection equipment, and to identify, predict and monitor the type, scale, distribution and degree of hazard of safety hazards in the railway security area and surrounding environment, thereby providing a basis for remediation, disaster relief, emergency repairs and advance deployment, thereby improving the overall management level of the railway surrounding environment.

[0063] Furthermore, the remote sensing method for safety hazards along the railway based on geographic information proposed in the present invention can be used to conduct regular revisit monitoring to overcome the problem of insufficient accuracy in the target feature extraction of safety hazards in the surrounding environment along the railway in the prior art, and to accurately identify the hidden dangers in the surrounding environment of the railway that affect railway driving safety.

[0064] By adopting the embodiments of the invention, it is possible to realize multi-temporal display of remote sensing images along the surrounding area, time series retrieval of spatial information of surrounding objects, positioning and retrieval of information on surrounding environmental changes, and to achieve functions such as grading and classification analysis of dangerous hazards in security areas.

[0065] In a specific embodiment of the present invention, in order to achieve multi-source remote sensing identification and change results to more effectively adapt to the subsequent data analysis of railway surrounding environment management and daily work applications of inspection personnel, the results are effectively integrated, encoded and aligned with the railway geographic information system.

[0066] The steps to effectively fuse, encode and register the results with the railway GIS include:

[0067] (1) Data preparation and preprocessing. Integrate the remote sensing identification and change results of various types of safety hazards in the railway surrounding environment; call the railway-specific geographic information data of the corresponding lines in the railway geographic information system, including vector data such as railway lines and important infrastructure, and perform geocoding and registration. Accurately geocode and register the remote sensing identification and change results with the spatial data in the railway geographic information system, generate a unique number for each result, and ensure the consistency of spatial position.

[0068] (2) Classification of safety hazards. Through the spatial feature extraction method, the registered remote sensing recognition and change results are scanned, and the safety hazard targets (such as dust screens, illegal buildings, etc.) are graded and classified according to the severity and urgency according to the railway application requirements.

[0069] (3) Vector data processing and conversion. The vector data of surface elements such as dust screens and other surface elements identified and changed by remote sensing are converted from the original coordinate system to the CGCS2000 (China Geodetic Coordinate System 2000) geographic coordinate system in order to match the positioning and analysis of the entire railway line.

[0070] (4) Spatial statistical analysis and parameter extraction. GIS spatial statistical analysis tools are used to extract parameters from remote sensing identification and change results, including location information, feature information (such as shape, size, color, etc.) and related image data, calculate the central longitude and latitude coordinates of each surface element, calculate the area information of each surface element to support further regional analysis, calculate the shortest distance of each surface element from the high-speed rail trunk line, and evaluate its potential impact on railway operation safety.

[0071] (5) Data integration and storage. Expand the attribute list of remote sensing identification and change results, and add the calculated location information, area, shortest distance and other data, as well as the original feature information, to the attribute list of the result remote sensing identification and change vector data. Export the vector data containing all necessary information, together with the number, into Excel format to facilitate subsequent statistical analysis, field verification and data integration.

[0072] By adopting the above-mentioned embodiments of the invention, it is possible to effectively realize the deep integration of multi-source remote sensing images and railway geographic information systems, improve the efficiency and accuracy of safety hazard detection and identification, and provide strong guarantees for the safe operation of railways.

[0073] Corresponding to the above method, the present invention also provides a remote sensing monitoring system for safety hazards along railways based on geographic information, the system comprising a computer device, the computer device comprising a processor and a memory, the memory storing computer instructions, the processor being used to execute the computer instructions stored in the memory, and when the computer instructions are executed by the processor, the system implements the steps of the method described above.

[0074] The method and system for remote sensing monitoring of safety hazards along railways based on geographic information proposed in the present invention can identify safety hazards based on multi-temporal remote sensing image data acquired for different safety hazard targets using a pre-trained generative adversarial network, and can monitor safety hazards in the surrounding environment of the railway based on multi-temporal remote sensing image data by integrating the safety hazard identification results identified in different phases.

[0075] The key points and advantages of the method and system for remote sensing monitoring of potential safety hazards along railways based on geographic information proposed in the present invention are:

[0076] (1) Geocoding and registration can be used to uniformly manage time and space dimensions, thereby solving the spatial correlation problem between safety hazards and railway equipment and facilities.

[0077] (2) It can utilize unified intelligent management of time / space dimensions through geocoding and registration, thereby resolving the spatial correlation problem between safety hazards and railway equipment and facilities.

[0078] (3) Generative adversarial networks can be used to distinguish between remote sensing image data with safety hazard elements and remote sensing image data without safety hazard elements. The proposed sample library annotation and augmentation method based on the transfer learning algorithm can solve the problem of the small number of data sets on environmental hazards around railways, so that the model trained based on this method can cover all types of safety hazards along the railway.

[0079] (4) The constructed safety hazard element target hazard sample library can be used to train and optimize the network. The deep learning network uses labeled data for repeated training to achieve a good discriminant model for change detection, which can effectively reduce the dependence on data volume and improve network robustness.

[0080] (5) It can realize multi-temporal display of remote sensing images along the route, time series retrieval of spatial information of surrounding objects, positioning and retrieval of information on surrounding environmental changes, and can be used for grading and classification analysis of dangerous hazards in security areas.

[0081] (6) It can effectively realize the deep integration of multi-source remote sensing images and railway geographic information systems, improve the efficiency and accuracy of safety hazard detection and identification, and provide strong guarantees for the safe operation of railways.

[0082] See also Figure 2 The computer device 00 includes: a processor 01, a memory 02, and a computer program stored in the memory 02 and executable on the processor 01. When the processor 01 executes the computer program, the human factor data server access control method provided in the above method embodiment is implemented.

[0083] Among them, the processor 01 is connected to the memory 02, such as through a bus 03. The processor 01 can be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It can implement or execute various exemplary logic blocks, modules and circuits described in combination with the contents disclosed in this application. The processor 01 can also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of DSP and microprocessors, etc. The bus 03 may include a path to transmit information between the above components. The bus 03 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus 130 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 2Only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. The memory 02 is used to store a computer program corresponding to the human factor data server access control method of the above embodiment of the present application, and the computer program is controlled and executed by the processor 01. The processor 01 is used to execute the computer program stored in the memory 02 to implement the content shown in the above method embodiment.

[0084] The embodiment of the present invention also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps of the aforementioned edge computing server deployment method are implemented. The computer-readable storage medium can be a tangible storage medium, such as a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a floppy disk, a hard disk, a removable storage disk, a CD-ROM, or any other form of storage medium known in the technical field.

[0085] It should be understood by those skilled in the art that the exemplary components, systems and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software or a combination of the two. Whether it is performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present invention are programs or code segments used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link via a data signal carried in a carrier.

[0086] It should be clear that the present invention is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present invention.

[0087] In the present invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with features of other embodiments or replace features of other embodiments.

[0088] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the embodiments of the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for remote sensing monitoring of safety hazards along railways based on geographic information, characterized in that: The method comprises the following steps: For predetermined safety hazard targets in the surrounding environment of the railway, remote sensing image data related to each safety hazard target at different times and / or from different data sources are obtained; wherein the remote sensing image data has been pre-processed including correction processing and standardization processing, and the label of each remote sensing image data includes geographic coding and registration results based on railway-specific geographic information; A pre-trained generative adversarial network for identifying safety hazards in the railway surrounding environment is used to extract features from each remote sensing image data and generate safety hazard identification results; wherein the generative adversarial network is trained using a pre-constructed safety hazard element target hazard sample library as real samples, and the safety hazard element target hazard sample library contains remote sensing image data and corresponding target hazard information; By integrating the safety hazard identification results obtained from remote sensing image data at different times related to various safety hazard targets, and combining the geocoding and registration results of the geographic information contained in the labels of the remote sensing image data, safety hazard monitoring of the railway surrounding environment based on multi-temporal remote sensing image data can be achieved.

2. The method according to claim 1, characterized in that The preprocessing of the remote sensing image data also includes multiple processings such as data mosaicking, bad pixel repairing, color uniformity and image cropping. The correction processing includes radiation correction and orthorectification. The standardization processing includes unifying the feature quantities of different scales carried in the remote sensing image data to the same scale.

3. The method according to claim 1, characterized in that: The method further comprises the step of fusing the remote sensing image data with the geographic information data, calling the railway-specific geographic information of the railway geographic information system to label the remote sensing image data with a label containing the geographic coding and the registration result, the step comprising: The vector data contained in the remote sensing image data of each safety hazard target is converted into one or more of the WGS84 geographic coordinate system, CGCS2000 geographic coordinate system, Albers plane coordinate system and UTM projection, and the railway-specific geographic information of the railway geographic information system is called for annotation, and the unique identifier of the geographic element is used as the annotation of the geographic coding and alignment results of each remote sensing image data.

4. The method according to claim 3, characterized in that The step of implementing the safety hazard monitoring of the railway surrounding environment based on multi-temporal remote sensing image data by combining the geocoding and registration results of the geographic information contained in the label of the remote sensing image data includes: Based on the geocoding and registration results of the geographic information contained in the labels of the remote sensing image data, the positions of the safety hazard targets corresponding to the remote sensing image data with safety hazards in the surrounding environment of the railway are determined in the WGS84 geographic coordinate system, CGCS2000 geographic coordinate system, Albers plane coordinate system or UTM projection, and based on the statistics of the multi-temporal remote sensing image data, the multi-temporal remote sensing image visualization display around the railway is realized, and based on the statistics of the multi-temporal remote sensing image data, the time series retrieval of the spatial information of the ground objects in the surrounding environment of the railway and the positioning and retrieval of the environmental change information are realized.

5. The method according to claim 1, characterized in that The method further comprises the step of pre-building a target potential hazard sample library of potential safety hazards, which step comprises: Perform image cropping on the remote sensing image data corresponding to each potential safety hazard target; Receive the target hidden danger information annotated by the personnel on the remote sensing image data after different cropping; Based on the remote sensing image data and target hidden danger information annotated by personnel, a sample label transfer network model is constructed; wherein the sample label transfer network model is used to generate corresponding safety hidden danger identification results based on the input remote sensing image data.

6. The method according to claim 5, characterized in that The target hidden danger information includes a safety hidden danger category and a safety hidden danger level, and the safety hidden danger identification result includes a safety hidden danger category and a safety hidden danger level.

7. The method according to claim 1, characterized in that The method also includes a pre-training step of generating an adversarial network, the step comprising: A pre-constructed target hidden danger sample library of safety hidden danger elements is used as a real sample, and a generator is used to generate false samples. The discriminator is trained and improved in an iterative process to distinguish between the real samples and the false samples, thereby realizing the pre-training of the generative adversarial network.

8. A remote sensing monitoring system for potential safety hazards along railways based on geographic information, comprising a processor, a memory and a computer program / instruction stored in the memory, characterized in that: The processor is used to execute the computer program / instructions, and when the computer program / instructions are executed, the system implements the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method as claimed in any one of claims 1 to 7 are implemented.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images