CN117709580A - Ocean disaster-bearing body vulnerability evaluation method based on SETR and geographic grid - Google Patents
Ocean disaster-bearing body vulnerability evaluation method based on SETR and geographic grid Download PDFInfo
- Publication number
- CN117709580A CN117709580A CN202311606866.2A CN202311606866A CN117709580A CN 117709580 A CN117709580 A CN 117709580A CN 202311606866 A CN202311606866 A CN 202311606866A CN 117709580 A CN117709580 A CN 117709580A
- Authority
- CN
- China
- Prior art keywords
- disaster
- space
- grid
- bearing body
- ocean
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000011156 evaluation Methods 0.000 title claims abstract description 54
- 238000000034 method Methods 0.000 claims abstract description 50
- 230000011218 segmentation Effects 0.000 claims abstract description 18
- 238000012545 processing Methods 0.000 claims abstract description 13
- 238000005516 engineering process Methods 0.000 claims abstract description 9
- 238000004364 calculation method Methods 0.000 claims description 44
- 230000006378 damage Effects 0.000 claims description 27
- 230000008569 process Effects 0.000 claims description 17
- 239000013598 vector Substances 0.000 claims description 17
- 238000010606 normalization Methods 0.000 claims description 12
- 238000011084 recovery Methods 0.000 claims description 11
- 238000010586 diagram Methods 0.000 claims description 10
- 238000004458 analytical method Methods 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- 239000000284 extract Substances 0.000 claims description 5
- 230000007246 mechanism Effects 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 abstract description 7
- 238000000605 extraction Methods 0.000 abstract description 7
- 238000012544 monitoring process Methods 0.000 abstract description 4
- 230000002265 prevention Effects 0.000 abstract description 4
- 230000009467 reduction Effects 0.000 abstract description 4
- 238000011835 investigation Methods 0.000 abstract description 3
- 238000013210 evaluation model Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 7
- 238000011160 research Methods 0.000 description 5
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000020477 pH reduction Effects 0.000 description 1
- 238000001556 precipitation Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 241000894007 species Species 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
- G06N3/0455—Auto-encoder networks; Encoder-decoder networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/05—Geographic models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/20—Finite element generation, e.g. wire-frame surface description, tesselation
- G06T17/205—Re-meshing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- Tourism & Hospitality (AREA)
- Geometry (AREA)
- General Health & Medical Sciences (AREA)
- Computer Graphics (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Entrepreneurship & Innovation (AREA)
- Computational Linguistics (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Primary Health Care (AREA)
- Remote Sensing (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a sea disaster-bearing body vulnerability evaluation method based on SETR and a geographic grid, which belongs to the technical field of intelligent sea data processing, and comprises the following steps: establishing a sample library, constructing a semantic segmentation model of the ocean disaster-bearing body remote sensing image based on an SETR network, constructing a grid space-time data set by using a geographic grid subdivision technology, establishing a vulnerability assessment method of the disaster-bearing body of grid and space value density, calculating a vulnerability index of the ocean disaster-bearing body, and realizing vulnerability assessment of the ocean disaster-bearing body in the area. Aiming at the problems of difficult investigation and extraction of massive marine disaster-bearing bodies, lack of intelligent early warning monitoring technology and the like, a marine disaster-bearing body vulnerability evaluation model based on SETR and geographic grids is constructed by utilizing the technologies of deep learning, geographic grid subdivision, space value density, disaster science and the like, intelligent and quick evaluation of the marine disaster-bearing bodies in a large-scale scene is realized, and the method has great significance for intelligent disaster monitoring early warning and disaster prevention and disaster reduction fine management of the marine disaster-bearing bodies under typhoon disasters.
Description
Technical Field
The invention relates to the technical field of intelligent ocean data processing, in particular to an ocean disaster-bearing body vulnerability evaluation method based on SETR and geographic grids.
Background
In recent years, global ocean disasters such as hurricanes, tsunamis, ocean acidification, sea level elevation and the like have caused significant impact on coastal areas. The background of the vulnerability evaluation and research of the ocean disaster-bearing body is to deeply understand the vulnerability of the ocean system to the disaster, and can provide scientific basis to support relevant decision making, so that the damage of the ocean disaster to the human society is reduced. The vulnerability of the ocean disaster-bearing body is evaluated, and the natural characteristics, ecological functions and sensitivity and resistance of human activities of the ocean system to the ocean disaster can be deeply known. This helps to take relevant action before the disaster occurs, reducing the risk of the disaster. The ocean disaster-bearing body vulnerability evaluation research basis can provide scientific analysis and prediction capability and decision support for disaster prevention and reduction. The development of vulnerability assessment of the ocean disaster bearing body is continuously promoted by continuous technical progress and research innovation.
Because of various kinds of ocean disaster-bearing bodies, including houses, roads, street lamps, communication facilities, seawalls, ocean cultivation facilities, coastal villages and towns and the like, the characteristic data of the ocean disaster-bearing bodies cannot be comprehensively found out; and the data related to the vulnerability evaluation of the ocean disaster-bearing body is wide and huge, such as mass space-time data of weather, hydrology, basic geography, human economy, traffic, remote sensing and the like, and the rapid and effective data organization and data analysis processing are difficult to realize by the traditional technical means. At present, the ocean management department is difficult to evaluate and analyze the ocean disaster-bearing body and finely manage ocean disaster prevention and reduction. Therefore, it is necessary to design a sea disaster-bearing body vulnerability evaluation method based on SETR and geographic grids.
Disclosure of Invention
The invention aims to provide a sea disaster-bearing body vulnerability evaluation method based on SETR and geographic grids, which solves the technical problems that the existing investigation and extraction of mass sea disaster-bearing bodies are difficult and the intelligent early warning and monitoring technical means are lacking.
Deep learning is a powerful machine learning technology, and can extract key features of a marine disaster-bearing body through learning related data of the marine disaster-bearing body and establish an accurate marine disaster-bearing body extraction model. The deep learning algorithm is utilized to realize large-scale and rapid extraction of ocean disaster bearing body targets from massive remote sensing data, and the evaluation efficiency and accuracy are improved. Geographic grid is a method of dividing geographic information into equally sized grid cells. The vulnerability assessment method based on the geographic grids can comprehensively utilize multi-source data including oceans, weather, geography, artificial factors and the like, and the multi-dimensional information of the ocean disaster bearing body is integrated into an assessment model through a deep learning algorithm, so that the comprehensiveness and the comprehensiveness of assessment are improved. The geographic grid may be adjusted according to the evaluation scale. The deep learning and geography grid-based method can evaluate the fine granularity according to the need, for example, the vulnerability of the ocean disaster bearing body in a local area can be evaluated, and the whole evaluation can be performed in a large area. Such flexibility enables the assessment method to accommodate assessment requirements of different scales.
Based on the vulnerability evaluation result, the disaster risk faced by the ocean system is evaluated, including disaster probability, potential loss and the like. And providing suggestions related to risk management and disaster response for a decision maker by using methods such as multi-criterion decision analysis, risk management framework and the like. The vulnerability assessment of the ocean disaster-bearing body involves knowledge of a plurality of disciplines, and cooperation and comprehensive analysis of interdisciplinary experts are needed. By fusing data, models, and methods of different disciplines, the vulnerability of the marine system can be more fully understood and assessed.
In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:
a sea disaster-bearing body vulnerability evaluation method based on SETR and geographic grids comprises the following steps:
s1: establishing a semantic segmentation sample library of the marine disaster-bearing body, constructing a semantic segmentation model of the marine disaster-bearing body remote sensing image based on an SETR network, extracting and classifying the marine disaster-bearing body in the remote sensing image, and calculating the geometric properties of the marine disaster-bearing body by a statistical pixel method;
s2: dividing space-time grids, constructing a grid space-time data set, and performing space-time correlation on massive space-time data by using a geographic grid subdivision technology to organize the data set with the grids as units;
s3: modeling regional geographic environment by using an oblique photogrammetry technology, classifying and extracting geographic element information, establishing a grid and space value density vulnerability assessment method of a disaster-bearing body, and constructing a marine disaster-bearing body vulnerability geographic calculation model to realize quantitative calculation of physical vulnerability of natural disasters;
s4: and calculating the vulnerability index of the ocean disaster-bearing body, and evaluating the vulnerability of the ocean disaster-bearing body in the region according to the vulnerability index of the disaster-bearing body.
Further, the specific process of constructing the ocean disaster-bearing body semantic segmentation model facing the remote sensing image in the step 1 is as follows:
(1) Serializing an input image into a one-dimensional vector, dividing the image into image blocks with the same size, encoding the spatial information of each image block to form final serial input, and adjusting each image block into a one-dimensional vector form;
(2) Inputting the serialized one-dimensional vector to a transducer encoder for extracting characteristics, wherein the transducer encoder comprises a plurality of transducer modules, and each module comprises a plurality of head self-attention modules, a normalization layer and a plurality of sensor layers;
(3) the output of the transducer encoder is passed to a decoder, which also consists of several transducer modules, which combines global context information with local details by using self-attention and cross-attention mechanisms, maps the encoded vectors back to image blocks, classifies at the pixel level, and generates a prediction of semantic segmentation.
Further, in the step 2, when the time-space correlation is performed on the massive time-space data, a geogrid code is generated, and the specific process is as follows:
(1) Acquiring the longitude and latitude heights and time nodes related to data;
(2) Expressing the longitude and latitude height as the form of integer division and second multiplied by 2048, and converting the longitude and latitude height into binary numbers; performing bit-by-bit crossing operation to form binary one-dimensional codes to form space codes in sea area information grid codes;
(3) Forming a time code by the time node according to the form of year, month and day;
(4) Combining the space code and the time code to form a body code;
(5) And generating check codes by the body codes to form sea area information grid codes in a unified way.
Further, the specific process of constructing the ocean disaster-bearing body vulnerability geographic calculation model in the step 3 is as follows:
(1) Constructing a rating index system, selecting disaster-bearing body elements to establish a space value density evaluation index system according to the geographical condition of an demonstration area and the exposure degree in ocean disasters, and determining the weight of each factor by expert scoring and normalization;
(2) Establishing a quantitative calculation model of vulnerability of a disaster-bearing body, carrying out geographic subdivision on an demonstration area according to 10m grids in the quantitative calculation model, dividing the demonstration area according to the subdivision grids, respectively calculating a value index and a damage index of each grid, and carrying out product finding:
V＝E×D (1)
wherein: v vulnerability index, E value index, D vulnerability index;
(3) Calculating value indexes, namely calculating the value quantity Q of each ground feature element in each grid, carrying out normalization processing on the calculated results, and then carrying out weighted summation, wherein the calculation formula is as follows:
wherein, E value density, Q value quantity of a certain ground object element in square kilometer of unit area, W ground object element weight, i represents the ith ground object element, n represents n ground object elements in total;
(4) The vulnerable index is calculated, wherein the vulnerable index of the evaluation unit is determined through two factors, namely a disaster risk level L and a ground object damage probability P of the evaluation unit, the L expresses the disaster risk of the evaluation unit caused by the difference of ground elevation, and the P is the statistical value of the damage probability of various ground objects in the whole evaluation unit, and the calculation formula is as follows:
D＝L×P (3)
wherein D is vulnerability index, L is disaster risk level, P is ground object damage probability, and the value range is 0 to 1.
Further, the calculation process of the ocean disaster-bearing body vulnerability index in the step 4 is as follows:
(1) Calculating the fragile index of the marine disaster-bearing body by taking the grid as a unit according to the value index and the fragile index of the single grid;
(2) And (5) summarizing the results of all grids to form the vulnerability evaluation of the ocean disaster-bearing body in the whole area.
Further, in the step 4, the calculation process of the ocean disaster-bearing body fragile index is as follows:
(1) Calculating the fragile index of the marine disaster-bearing body by taking the grid as a unit according to the value index and the fragile index of the single grid;
(2) And (5) summarizing the results of all grids to form the vulnerability evaluation of the ocean disaster-bearing body in the whole area.
6. The ocean disaster recovery system vulnerability assessment method based on SETR and geographic grids, according to claim 2, is characterized in that: the specific process of serializing the input image is as follows:
evenly dividing pictures intoWhere H is the picture height and W is the picture width, then flattening each image block into a one-dimensional vector sequence, mapping the vectorized image block p to the C dimension using linear mapping, i.eC is the size of the hidden channel, is the size of the class, will be for eachThe pixel is subjected to position coding to obtain p i Then and vector e i Adding to obtain the final input sequence E= { E 1 +p 1 ，e 2 +p 2 ，...，e L +p L Sequence length L is->The final dimension to be input to the transducer encoder is +.>e 1 ...e L Representing the corresponding vector of each image block, p 1 ...p L Representing the corresponding position information code for each image block.
Further, the decoder performs pixel level segmentation on the image, extracts a feature from the encoder every 6 layers, and then extracts the feature from the encoderIs adjusted to->Is processed by a three-layer convolution network of 1×1,3×3 and 3×3 respectively, wherein the first layer and the third layer reduce the channel number to half and outputUp-sampling by 4 times by bilinear interpolation to obtain +.>Four->The features of the top layer and the three fused features are then spliced to obtain +.>In the size of (2)And after adding the functions element by element, applying additional 3×3 convolution, and finally obtaining a final H×W×C feature map through 4 times up-sampling and convolution, and carrying out normalization processing on the class probability of the pixel points by adopting a softmax function to obtain a final semantic segmentation result.
Further, the specific process of space-time grid division in step S2 is as follows:
the subdivision space of the earth surface space grid subdivision adopts a longitude and latitude coordinate space, in order to ensure the division of longitude and latitude height, integer division and integer second, the earth space is expanded into a longitude and latitude space of 2 integer power, the longitude and latitude space is defined as 512 degrees multiplied by 512 degrees grid, 60 'space of each degree is expanded to 64', 60 'space of each minute is expanded to 64', the subdivision space is divided to 32 stages step by step according to an octave method, the minimum expression precision can be accurate to 1.5 cm, the part in the 180 degrees multiplied by 360 degrees range in the expanded longitude and latitude coordinate space is consistent with the actual geographic space, the part in the range exceeding 180 degrees multiplied by 360 degrees has no actual geographic meaning, when the elevation is fixed as the earth surface height, the land and sea space-time information grid is taken as the earth plane grid, and when the elevation value is not needed, the plane grid can be used.
Further, the ground object damage probability calculation process is as follows:
the calculation method of the damage probability of the unit ground features is evaluated, the types of the ground features, the parameters of the ground features and the area ratio occupied by various ground features in the unit grid are determined, weighted summation is carried out, the calculation formula is as follows, and the normalization processing is carried out on the result after calculation, so that the value range of the damage probability P of the ground features is 0-1;
wherein P is the damage probability of the ground object of the evaluation unit, F is the damage probability of the single ground object, S is the area percentage of the single ground object in the evaluation unit, and since the evaluation unit is a kilometer grid, the area of each grid is 1KM 2 The actual area of each ground object is the area percentage of the ground object in the evaluation unit.
Further, the specific process of disaster risk classification is as follows:
the calculation of disaster risk level takes kilometer grids as units, calculates the average ground elevation in each kilometer grid, divides the average ground elevation into 10 grades, is Gao Chengyue low, and is characterized in that the larger the disaster risk is, the larger the area with the elevation exceeding 180 meters is not affected by ocean disasters according to historical ocean disaster statistics data, the disaster risk level calculation of ground feature elements takes the space object of the ground feature elements as a unit, and a plurality of space diagram layers are needed to be stacked for space analysis calculation so as to determine the disaster risk level formed by space position distribution.
Due to the adoption of the technical scheme, the invention has the following beneficial effects:
(1) In the aspect of disaster-bearing body extraction, the invention realizes large-scale and rapid extraction of ocean disaster-bearing body data through remote sensing data and SETR network, and improves the efficiency and saves the cost compared with the traditional manual sketching method;
(2) In the aspect of marine disaster-bearing body evaluation, the correlation analysis of multi-source heterogeneous data is realized through a global geographic subdivision grid, the space-time correlation and the internal value of each disaster-causing factor data are fully mined by utilizing the same space-time attribute of different format and different category data in the same grid, the limitation of the vulnerability knowledge of the disaster-bearing bodies among different types is reduced, and the scientificity and reliability of an evaluation result are improved;
(3) The deep learning model can provide scientific basis for the establishment of emergency measures, namely is favorable for guiding the practice of emergency management, reduces the harm of disaster events to the greatest extent, and improves the effectiveness and scientificity of the emergency management.
Drawings
FIG. 1 is a flow chart of a method for evaluating marine disaster recovery bodies according to the present invention;
FIG. 2 is a schematic diagram of a transducer-based encoder;
FIG. 3 is a schematic diagram of multi-layer feature fusion decoding;
FIG. 4 is a schematic floor plan;
FIG. 5 is a schematic diagram of traffic lane number calculation;
FIG. 6 is a schematic illustration of land-sea space-time geographic meshing;
FIG. 7 is a diagram of a global land-sea space-time grid coding structure;
FIG. 8 is a diagram of the effect of shoreline trellis encoding;
FIG. 9 is a diagram of the effect of sea-island trellis encoding;
FIG. 10 is a schematic diagram of vulnerability assessment of a marine disaster carrier based on a geographic grid.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described in detail below by referring to the accompanying drawings and by illustrating preferred embodiments. It should be noted, however, that many of the details set forth in the description are merely provided to provide a thorough understanding of one or more aspects of the invention, and that these aspects of the invention may be practiced without these specific details.
As shown in fig. 1, the ocean disaster-bearing body vulnerability evaluation method based on SETR and geographic grids comprises the following steps:
s1, establishing a semantic segmentation sample library, setting SETR network parameters, and constructing a semantic segmentation model of the ocean disaster-bearing body remote sensing image to extract, classify and calculate geometric attributes of the ocean disaster-bearing body in the remote sensing image. The specific implementation is as follows:
(1) And (5) image serialization processing. Evenly dividing pictures intoWherein H is the picture height and W is the picture width) and then flattening each image block into a one-dimensional vector sequence, mapping the vectorized image block p to the C dimension, i.e. & lt/EN & gt, using linear mapping>C is the size of the hidden channel, i.e., the size of the class; each pixel is subjected to position coding to obtain p i Then and vector e i Adding to obtain the final input sequence E= [ E 1 +p 1 ，e 2 +p 2 ，...，e L +p L Sequence length L is->The final dimension to be input to the transducer encoder is +.>e 1 ...e L Representing the corresponding vector of each image block, p 1 ...p L Representing the corresponding position information code for each image block.
(2) And (5) encoding a transducer. Features are learned using a one-dimensional sequence E as input and a transducer as encoder, each transducer layer having a global receptive field. Each transducer Layer consists of Multi-Head Self-Attention (MSA), layer Normalization (LN), and Multi-Layer Perceptron (MLP) layers, for a total of 24 layers. As shown in fig. 2.
Input to the self-attention module is a set of tuples, input isThe query, key, value is calculated as follows: query=z l-1 W Q ，key＝Z l-1 W K ，value＝Z l-1 W V WhereinD is the dimension of (query, key, value) for the learnable parameters.
Z l-1 The characteristics of the first layer of the transducer are shown, and l is the first layer of the transducer.
The self-attention module SA will do the following:
SA () represents the operation of the self-attention module, Z l-1 Characterizing the transformer layer 1, softmax is a normalized exponential function commonly used in deep learning,d is the dimension of (query, key, value) for the learnable parameters.
The multi-head self-attention module MSA is an extension of m independent self-attention operations and projects their serial outputs:
MSA(Z l-1 )＝[SA 1 (Z l-1 )；SA 2 (Z l-1 )；...；SA m (Z l-1 )]W O (5)
SA 1 ...SA m representing m self-attention module operations, MSA () represents a multi-head self-attention operation, mainly to enable encoding of image features.
Wherein the method comprises the steps ofd is usually set to +.>
The output of the MSA is converted by a remaining skipped MLP block, the layer output is:
MLP means that the multilayer perceptron carries out characteristic convolution coding processing on the image, and the whole function is to complete the coding processing of the first layer of transformers on the image characteristics.
Z l Representing the characteristics of the first layer of the transducer.
Finally, the output { Z of each layer of the transducer encoder is formed 1 ，Z 2 ，...，Z L }。
(3) The decoder is designed to perform pixel level segmentation on the image. Slave codeExtracting a feature every 6 layers in the device, and then extracting the feature from the deviceIs adjusted to->Is processed by a (1X 1, 3X 3) three-layer convolution network, wherein the first layer and the third layer reduce the channel number to half and outputUp-sampling by 4 times by bilinear interpolation to obtain +.>These four +.>The features of the top layer and the three fused features are then spliced to obtainAfter adding the function element by element, an additional 3 x 3 convolution is applied, and finally the final H x W x C feature map is obtained by 4 times up-sampling and convolution. And carrying out normalization processing on the class probability of the pixel points by adopting a softmax function to obtain a final semantic segmentation result. As shown in fig. 3 and 4.
(4) Based on the image semantic segmentation result, calculating the geometric properties of the ocean disaster-bearing body, such as the length, the area, the perimeter and the like of the disaster-bearing body by adopting a statistical pixel method.
S2, dividing the space-time grid, and constructing a grid data set. The land-sea space-time information grid takes the intersection point of the primary meridian and the equatorial plane of the reference ellipsoid as the origin, and the adopted longitude and latitude coordinate system and the earth reference ellipsoid follow the rule of GJB 6304-2008.
(1) And (5) space grid division. The subdivision space of the earth surface space grid subdivision adopts a longitude and latitude coordinate space, in order to ensure the division of the longitude and latitude height, the earth space is expanded into a longitude and latitude space of 2 integer power, the longitude and latitude space is defined as 512 degrees multiplied by 512 degrees grid, the 60 'space of each degree is expanded to 64', the 60 'space of each minute is expanded to 64', the earth surface space grid subdivision is divided to 32 stages according to an octave method (table 1), and the minimum expression precision can be accurate to 1.5 cm. For the part in the range of 180 degrees×360 degrees in the extended longitude and latitude coordinate space to coincide with the actual geographic space, the part exceeding the range of 180 degrees×360 degrees has no actual geographic meaning. When the elevation is fixed to be the surface elevation of the earth, the Guangxi land sea space-time information grid is the earth plane grid. When elevation values are not needed, the planar grid is used.
Table 1 statistics of the grid dimensions of each level
(2) And (5) grid coding. The global land-sea space-time grid code should be a feature combination code and be composed of 33-bit characters. The first 32 bits are the body code and the last bit is the check code. The body codes are arranged from left to right in sequence: 24-bit sea area information grid space coding section and 8-bit time coding section. For a given longitude and latitude height and time point of data, a global land-sea space-time grid code is generated as follows.
1) Acquiring the longitude and latitude heights and time nodes related to data;
2) The warp and weft heights are expressed in terms of integer fractions and seconds×2048, which are then converted to binary numbers. And performing bit-by-bit crossover operation to form binary one-dimensional codes, namely, forming space codes in the global land-sea space-time grid codes.
3) The time nodes are formed into time codes in the form of years, months and days.
4) And combining the space code and the time code to form the body code.
5) And generating check codes by the body codes to form the global land-sea space-time grid codes in a unified way.
For example, the longitude and latitude height of a given datum is 109.03 ° E, 21.56 ° N, 6372 km from the earth's center height (earth's surface height). Firstly, converting the height into corresponding degrees, wherein the calculation formula is as follows: height (unit: degree) =height (unit: kilometer)/40000 (equatorial circumference, unit: kilometer) ×360°, i.e., 57.348 ° H. Data represented by the degrees is converted into binary numbers. Taking longitude as an example, it is expressed in degrees in seconds: 109 deg. 1'48.0 ", for direct conversion of degree, score into binary number, the first bit of the integer represents positive and negative values, 9 bits total, 6 bits total. The method comprises the following steps of: (001101101) 2 、(000001) 2 The second and second fractions are multiplied by 2048 (minimum accuracy of the mesh division) and converted into binary numbers (11000000000000000) 2. Longitude is converted into (00110110100000111000000000000000) 2 . The latitude and the altitude are the same as each other, which are 00001010110000110010000000000000 2 And
(00011100101010011010011001100110) 2 . According to the bitwise interleaving method, a spatial encoding is formed. The values are taken according to the order of one bit of each warp and weft height, and finally the warp and weft heights are formed (000000100101011101110000111010001000001000110111101000011000000001001000000001001000000001001000) 2 The corresponding hexadecimal number is 025770E88237A18048048048. The time nodes of the data are represented as time codes in a year, month and day format. Such as 20170101. For missing time data, this segment is assigned a default value of 00000000. The spatial encoding and the temporal encoding are combined to form an ontology code and a check code is generated. And finally, coding is completed.
(3) A mesh dataset is constructed. Taking the geographic grids as space-time correlation, discretely loading marine environment data, marine carrier vulnerability data, forecast data of strong typhoon effect (such as wind intensity, wind direction, wind quantity and precipitation) and the like into each grid to form a grid data set.
S3, constructing a physical model for evaluating vulnerability of the ocean disaster-bearing body. Dividing the whole space into evaluation units by using grids, and calculating the sum of the ground feature value quantities in each grid to serve as a value index E (Economic value); meanwhile, according to the ocean disaster type, the Degree that the position of each grid unit is easily affected by the disaster is calculated and used as a vulnerability index D (easy-damage device), and the product of the two indexes is the vulnerability index V (Vulnerability) of the evaluation unit.
(1) And constructing a rating index system. 10 ground object elements are selected to establish a space value density evaluation index system, and the weight of each factor is determined through expert scoring, which is shown in table 2. The building is a ground feature and reflects the distribution of people, and in typhoon storm surge disasters, the building is not necessarily destroyed, but people in the building are evacuated, so that economic losses are formed, and the building has a large weight.
TABLE 2 ground object factor and weight
(2) And (5) calculating a value index. The value density E is calculated as follows:
1) Building price value density calculation. The building value density is the sum of the values of all the buildings in a unit area, and is calculated by the building area and unit price of the building. The parameters introduced include building floor area, floor number, price of different kinds of buildings. The floor area and the floor number are extracted from the urban three-dimensional model, and the building price refers to urban economic statistical data.
Wherein, S is the building bottom area; l building floor number; building price per unit area (square meter); each building within the i kilometer grid.
Buildings are classified into different types: greenhouse, neighborhood, common house, high-rise building, public facilities, etc., different building types, and p values are different. The layer number L of the building is calculated according to the height of the building, the layer height of a common house or a high-rise building is 3 meters, and the layer height of a neighborhood or a public place is 3.5 meters. As shown in fig. 4.
Table 3 building classification table
2) And calculating the road value density. Roads are an important infrastructure of cities, are easily damaged in typhoons and storm surge, and are also important evaluation factors. The road information is easy to obtain, and accurate quantitative calculation can be performed.
Wherein, L is the road length; w number of lanes; p value weight is determined according to road grade; each road within the i km grid.
Highway grades are generally classified as: expressway, primary road, secondary road, etc., and the p values of different road grades are different. According to the related design standard of the Chinese traffic road, each lane of the urban road is 3.5 meters wide, and according to the parameters, the number of lanes of the main road in the city can be calculated by combining the extracted road width information. As shown in fig. 5.
Table 4 road classification table
3) Population factors influence factor calculation. Population is an important factor in urban natural disaster vulnerability assessment. In many disaster loss evaluation researches at home and abroad, the death population is taken as an index, but the death population can only be obtained through post-disaster statistical data and is not easy to be taken as a vulnerability evaluation index. For vulnerability assessment requiring quantitative calculation, population factors are mainly reflected in the spatial distribution of population and the cost of large-scale personnel evacuation in face of disasters.
Q＝N×P (9)
Where N is the evacuation population, P is the cost required for evacuating 1 person, and the value of P is related to factors such as evacuation distance, evacuation settling time, and the like, and specific analysis is required.
4) And calculating the value density of other elements. Mainly comprises the following categories:
a. traffic facilities:
Q＝N×P (10)
wherein, the area of the S bottom surface, the number of L floors and the unit price of P (yuan/square meter);
b. linear water system:
wherein, the length of the river L and the corresponding value of the water quality of different types P are measured;
c. planar water system:
wherein, the area of the S water area and the corresponding value of the water quality of different types P are provided;
d. vegetation:
wherein, the S vegetation areas and the P are the corresponding value amounts of different vegetation types;
e. pipeline:
wherein, the length of the L pipeline and the corresponding value of the P pipelines of different types;
f. tourist attractions:
wherein, the S scenic spot area and the P are the corresponding value amounts of scenic spots of different types;
g. dykes and dams:
wherein, L dykes and dams length, H dykes and dams height, P dykes and dams cost per unit volume.
(3) And (5) calculating the vulnerability index, wherein the calculation comprises the division of the disaster risk level L and the calculation of the ground object damage probability P.
1) And (5) grading disaster risk. L is calculated by taking a kilometer grid as a unit, and the average ground elevation in each kilometer grid is calculated and divided into 10 grades, wherein the lower the Gao Chengyue is, the larger the disaster risk is. According to historical marine disaster statistics, areas with elevations exceeding 180 meters are generally not affected by marine disasters. The disaster risk level calculation of various ground object elements takes the space object of the ground object elements as a unit (such as a single building or a section of a road), and a plurality of space diagram layers are needed to be overlapped for space analysis calculation so as to determine the disaster risk level formed by the space position distribution.
TABLE 5 spatial grid cell hazard class
2) And (5) calculating the damage probability of the ground object. The calculation method for evaluating the damage probability of the unit ground features determines the types of the ground features in the unit grid, the main parameters of the ground features and the area ratio occupied by various ground features, and carries out weighted summation, wherein the calculation formula is as follows. And (3) carrying out normalization processing on the result after calculation to ensure that the value range of P is (0-1).
Wherein P is the damage probability of the ground object of the evaluation unit, F is the damage probability of the single ground object, and S is the area percentage of the single ground object in the evaluation unit. Since the evaluation unit is a kilometer grid, the area of each grid is 1KM 2 The actual area of each ground object is the area percentage of the ground object in the evaluation unit.
The damage probability F of each type of ground object is shown in Table 6. The building is the most important ground factor, and the vulnerability of different building types, floor heights and building ages is different.
TABLE 6 probability of damage to various terrain
3) And calculating the vulnerability index according to the formula (3).
S4, calculating the vulnerability index of the ocean disaster-bearing body. Combining DEM data, ocean disaster-bearing body data and storm tide value simulation results, performing hybrid calculation of spatial positions and attribute values, and calculating the submerged depth of each grid according to the difference value of tide level and ground elevation; calculating the value index and the easy-to-damage index of various elements in the grid by taking the grid as a unit; and (3) calculating the vulnerability index of each grid according to the formula (1), and then summarizing the results of all grids to form the vulnerability evaluation of the ocean disaster-bearing body of the whole area.
Aiming at the problems of difficult investigation and extraction of massive marine disaster-bearing bodies, lack of intelligent early warning monitoring technology and the like, the embodiment provides and realizes a marine disaster-bearing body vulnerability evaluation method based on SETR and geographic grid evaluation, and combines deep learning, disaster science, geography, marine science related to disaster species and the like to construct a marine disaster-bearing body vulnerability evaluation model. Experiments prove that the method can be used for disaster-bearing body response mechanism research, disaster-bearing body evaluation and disaster early warning, and provides basis for disaster prevention and reduction refined management in the coming time of strong typhoons.
The foregoing is merely a preferred embodiment of the present invention and it should be noted that modifications and adaptations to those skilled in the art may be made without departing from the principles of the present invention, which are intended to be comprehended within the scope of the present invention.
Claims (10)
1. A sea disaster-bearing body vulnerability evaluation method based on SETR and geographic grids is characterized by comprising the following steps: the method comprises the following steps:
s1: establishing a semantic segmentation sample library of the marine disaster-bearing body, constructing a semantic segmentation model of the marine disaster-bearing body remote sensing image based on an SETR network, extracting and classifying the marine disaster-bearing body in the remote sensing image, and calculating the geometric properties of the marine disaster-bearing body by a statistical pixel method;
s2: dividing space-time grids, constructing a grid space-time data set, and performing space-time correlation on massive space-time data by using a geographic grid subdivision technology to organize the data set with the grids as units;
s3: modeling regional geographic environment by using an oblique photogrammetry technology, classifying and extracting geographic element information, establishing a grid and space value density vulnerability assessment method of a disaster-bearing body, and constructing a marine disaster-bearing body vulnerability geographic calculation model to realize quantitative calculation of physical vulnerability of natural disasters;
s4: and calculating the vulnerability index of the ocean disaster-bearing body, and evaluating the vulnerability of the ocean disaster-bearing body in the region according to the vulnerability index of the disaster-bearing body.
2. The ocean disaster recovery system vulnerability assessment method based on SETR and geographic grids, according to claim 1, is characterized in that: the specific process for constructing the ocean disaster-bearing body semantic segmentation model facing the remote sensing image in the step 1 is as follows:
(1) Serializing an input image into a one-dimensional vector, dividing the image into image blocks with the same size, encoding the spatial information of each image block to form final serial input, and adjusting each image block into a one-dimensional vector form;
(2) Inputting the serialized one-dimensional vector to a transducer encoder for extracting characteristics, wherein the transducer encoder comprises a plurality of transducer modules, and each module comprises a plurality of head self-attention modules, a normalization layer and a plurality of sensor layers;
(3) the output of the transducer encoder is passed to a decoder, which also consists of several transducer modules, which combines global context information with local details by using self-attention and cross-attention mechanisms, maps the encoded vectors back to image blocks, classifies at the pixel level, and generates a prediction of semantic segmentation.
3. The ocean disaster recovery system vulnerability assessment method based on SETR and geographic grids, according to claim 2, is characterized in that: in the step 2, the geographic grid code is generated when the time-space correlation is carried out on massive time-space data, and the specific process is as follows:
(1) Acquiring the longitude and latitude heights and time nodes related to data;
(2) Expressing the longitude and latitude height as the form of integer division and second multiplied by 2048, and converting the longitude and latitude height into binary numbers; performing bit-by-bit crossing operation to form binary one-dimensional codes to form space codes in sea area information grid codes;
(3) Forming a time code by the time node according to the form of year, month and day;
(4) Combining the space code and the time code to form a body code;
(5) And generating check codes by the body codes to form sea area information grid codes in a unified way.
4. The ocean disaster recovery system vulnerability assessment method based on SETR and geographic grids, according to claim 1, is characterized in that: the specific process for constructing the ocean disaster-bearing body vulnerability geographic calculation model in the step 3 is as follows:
(1) Constructing a rating index system, selecting disaster-bearing body elements to establish a space value density evaluation index system according to the geographical condition of an demonstration area and the exposure degree in ocean disasters, and determining the weight of each factor by scoring and normalization;
(2) Establishing a quantitative calculation model of vulnerability of a disaster-bearing body, carrying out geographic subdivision on an demonstration area according to 10m grids in the quantitative calculation model, dividing the demonstration area according to the subdivision grids, respectively calculating a value index and a damage index of each grid, and carrying out product finding:
V＝E×D (1)
wherein: v vulnerability index, E value index, D vulnerability index;
(3) Calculating value indexes, namely calculating the value quantity Q of each ground feature element in each grid, carrying out normalization processing on the calculated results, and then carrying out weighted summation, wherein the calculation formula is as follows:
wherein, E value density, Q value quantity of a certain ground object element in square kilometer of unit area, W ground object element weight, i represents the ith ground object element, n represents n ground object elements in total;
(4) The vulnerable index is calculated, wherein the vulnerable index of the evaluation unit is determined through two factors, namely a disaster risk level L and a ground object damage probability P of the evaluation unit, the L expresses the disaster risk of the evaluation unit caused by the difference of ground elevation, and the P is the statistical value of the damage probability of various ground objects in the whole evaluation unit, and the calculation formula is as follows:
D＝L×P (3)
wherein D is vulnerability index, L is disaster risk level, P is ground object damage probability, and the value range is 0 to 1.
5. The ocean disaster recovery system vulnerability assessment method based on SETR and geographic grids, according to claim 1, is characterized in that: the calculating process of the ocean disaster-bearing body fragile index in the step 4 is as follows:
(1) Calculating the fragile index of the marine disaster-bearing body by taking the grid as a unit according to the value index and the fragile index of the single grid;
(2) And (5) summarizing the results of all grids to form the vulnerability evaluation of the ocean disaster-bearing body in the whole area.
6. The ocean disaster recovery system vulnerability assessment method based on SETR and geographic grids, according to claim 2, is characterized in that: the specific process of serializing the input image is as follows:
evenly dividing pictures intoWherein H is the picture height and W is the picture width, then flattening each image block into a one-dimensional vector sequence, mapping the vectorized image block p to the C dimension, i.e. & lt/EN & gt, using linear mapping>C is the size of the hidden channel, is the size of the classification, and will perform position coding on each pixel to obtain p i Then and vector e i Adding to obtain the final input sequence E= { E 1 +p 1 ，e 2 +p 2 ，...，e L +p L Sequence length L is->The final dimension to be input to the transducer encoder is +.>e 1 ...e L Representing the corresponding vector of each image block, p 1 ...p L Representing the corresponding position information code for each image block.
7. The ocean disaster recovery system vulnerability assessment method based on SETR and geographic grids, according to claim 2, is characterized in that: the decoder performs pixel level segmentation on the image, extracts a feature from the encoder every 6 layers, and then extracts the feature from the encoderIs adjusted to->Is processed by a three-layer convolution network of 1 x 1,3 x 3 and 3 x 3, respectively, wherein the first layer and the third layer reduce the channel number to half and output +.>Up-sampling by 4 times by bilinear interpolation to obtain +.>Four->The features of the top layer and the three fused features are then spliced to obtain +.>After adding the function element by element, applying additional 3×3 convolution, finally obtaining a final H×W×C feature map through 4 times up-sampling and convolution, and normalizing the class probability of the pixel points by adopting a softmax function to obtain a final semantic segmentation result.
8. The ocean disaster recovery system vulnerability assessment method based on SETR and geographic grids, according to claim 1, is characterized in that: the specific process of space-time grid division in the step S2 is as follows:
the subdivision space of the earth surface space grid subdivision adopts a longitude and latitude coordinate space, in order to ensure the division of longitude and latitude height, integer division and integer second, the earth space is expanded into a longitude and latitude space of 2 integer power, the longitude and latitude space is defined as 512 degrees multiplied by 512 degrees grid, 60 'space of each degree is expanded to 64', 60 'space of each minute is expanded to 64', the subdivision space is divided to 32 stages step by step according to an octave method, the minimum expression precision can be accurate to 1.5 cm, the part in the 180 degrees multiplied by 360 degrees range in the expanded longitude and latitude coordinate space is consistent with the actual geographic space, the part in the range exceeding 180 degrees multiplied by 360 degrees has no actual geographic meaning, when the elevation is fixed as the earth surface height, the land and sea space-time information grid is taken as the earth plane grid, and when the elevation value is not needed, the plane grid can be used.
9. The ocean disaster recovery system vulnerability assessment method based on SETR and geographic grids, according to claim 4, is characterized in that: the ground object damage probability calculation process comprises the following steps:
the calculation method of the damage probability of the unit ground features is evaluated, the types of the ground features, the parameters of the ground features and the area ratio occupied by various ground features in the unit grid are determined, weighted summation is carried out, the calculation formula is as follows, and the normalization processing is carried out on the result after calculation, so that the value range of the damage probability P of the ground features is 0-1;
wherein P is the damage probability of the ground object of the evaluation unit, F is the damage probability of the single ground object, S is the area percentage of the single ground object in the evaluation unit, and since the evaluation unit is a kilometer grid, the area of each grid is 1KM 2 The actual area of each ground object is the area percentage of the ground object in the evaluation unit.
10. The ocean disaster recovery system vulnerability assessment method based on SETR and geographic grids, according to claim 4, is characterized in that: the specific process of disaster risk classification is as follows:
the calculation of disaster risk level takes kilometer grids as units, calculates the average ground elevation in each kilometer grid, divides the average ground elevation into 10 grades, is Gao Chengyue low, and is characterized in that the larger the disaster risk is, the larger the area with the elevation exceeding 180 meters is not affected by ocean disasters according to historical ocean disaster statistics data, the disaster risk level calculation of ground feature elements takes the space object of the ground feature elements as a unit, and a plurality of space diagram layers are needed to be stacked for space analysis calculation so as to determine the disaster risk level formed by space position distribution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311606866.2A CN117709580A (en) | 2023-11-29 | 2023-11-29 | Ocean disaster-bearing body vulnerability evaluation method based on SETR and geographic grid |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311606866.2A CN117709580A (en) | 2023-11-29 | 2023-11-29 | Ocean disaster-bearing body vulnerability evaluation method based on SETR and geographic grid |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117709580A true CN117709580A (en) | 2024-03-15 |
Family
ID=90150657
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311606866.2A Pending CN117709580A (en) | 2023-11-29 | 2023-11-29 | Ocean disaster-bearing body vulnerability evaluation method based on SETR and geographic grid |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117709580A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118135424A (en) * | 2024-05-06 | 2024-06-04 | 中科星图智慧科技有限公司 | Ecological environment supervision method and device based on remote sensing image and GIS |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112835997A (en) * | 2020-10-22 | 2021-05-25 | 北京旋极伏羲科技有限公司 | Power grid inspection space grid coding method based on GeoSOT grid |
| CN113688813A (en) * | 2021-10-27 | 2021-11-23 | 长沙理工大学 | Multi-scale feature fusion remote sensing image segmentation method, device, equipment and storage |
| CN115018285A (en) * | 2022-05-25 | 2022-09-06 | 南宁师范大学 | Storm surge and sea wave fine early warning system and early warning method |
| US20230090423A1 (en) * | 2021-09-18 | 2023-03-23 | Institute Of Mountain Hazards And Environment, Chinese Academy Of Sciences | Method for whole-process numerical simulation and hazard forecast of mountain disaster |
| CN116152678A (en) * | 2022-12-01 | 2023-05-23 | 广西科学院 | Marine disaster-bearing body identification method based on twin neural network under small sample condition |
| CN117095287A (en) * | 2023-07-27 | 2023-11-21 | 浙江大学 | Remote sensing image change detection method based on space-time interaction transducer model |
-
2023
- 2023-11-29 CN CN202311606866.2A patent/CN117709580A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112835997A (en) * | 2020-10-22 | 2021-05-25 | 北京旋极伏羲科技有限公司 | Power grid inspection space grid coding method based on GeoSOT grid |
| US20230090423A1 (en) * | 2021-09-18 | 2023-03-23 | Institute Of Mountain Hazards And Environment, Chinese Academy Of Sciences | Method for whole-process numerical simulation and hazard forecast of mountain disaster |
| CN113688813A (en) * | 2021-10-27 | 2021-11-23 | 长沙理工大学 | Multi-scale feature fusion remote sensing image segmentation method, device, equipment and storage |
| CN115018285A (en) * | 2022-05-25 | 2022-09-06 | 南宁师范大学 | Storm surge and sea wave fine early warning system and early warning method |
| CN116152678A (en) * | 2022-12-01 | 2023-05-23 | 广西科学院 | Marine disaster-bearing body identification method based on twin neural network under small sample condition |
| CN117095287A (en) * | 2023-07-27 | 2023-11-21 | 浙江大学 | Remote sensing image change detection method based on space-time interaction transducer model |
Non-Patent Citations (3)
| Title |
|---|
| 田振坤;傅莺莺;刘素红;: "基于异构机器学习算法融合的遥感影像分类", 计算机科学, no. 05, 15 May 2019 (2019-05-15), pages 235 - 240 * |
| 程承旗，曲腾腾等: "数字孪生城市空间网格框架与GDS数据平台技术研究", 《信息通信技术与政策》, no. 3, 31 March 2020 (2020-03-31), pages 1 - 8 * |
| 邬满，刘金等: "基于深度学习的海洋承灾体评价方法研究", 《测绘地理信息》, vol. 47, no. 6, 31 December 2022 (2022-12-31), pages 91 - 95 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118135424A (en) * | 2024-05-06 | 2024-06-04 | 中科星图智慧科技有限公司 | Ecological environment supervision method and device based on remote sensing image and GIS |
| CN118135424B (en) * | 2024-05-06 | 2024-07-12 | 中科星图智慧科技有限公司 | Ecological environment supervision method and device based on remote sensing image and GIS |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Spatio-temporal urban growth dynamics of Lagos Metropolitan Region of Nigeria based on Hybrid methods for LULC modeling and prediction | |
| Pellicani et al. | Assessing landslide exposure in areas with limited landslide information | |
| Degerli et al. | Using the remote sensing method to simulate the land change in the year 2030 | |
| CN114240119A (en) | Digital twin-based flood control and waterlogging prevention system and early warning method for whole elements of territorial universe | |
| CN111340668B (en) | Typhoon disaster assessment system | |
| CN111859779B (en) | Method and device for early warning of third party construction damage risk of gas pipe network | |
| CN117709580A (en) | Ocean disaster-bearing body vulnerability evaluation method based on SETR and geographic grid | |
| Zhang et al. | Using street view images to identify road noise barriers with ensemble classification model and geospatial analysis | |
| Qiu et al. | Disaggregating population data for assessing progress of SDGs: methods and applications | |
| CN116151610B (en) | Method for simulating risk exposure space of disaster-bearing body on underlying surface of heterogeneous city | |
| Liu et al. | A quantitative method for storm surge vulnerability assessment–a case study of Weihai city | |
| CN117332909B (en) | Multi-scale urban waterlogging road traffic exposure prediction method based on intelligent agent | |
| Wang et al. | Risk assessment of typhoon disaster chains in the Guangdong–Hong Kong–Macau greater bay area, China | |
| Li et al. | Urban flood risk assessment based on DBSCAN and K-means clustering algorithm | |
| CN109784720A (en) | The associated power distribution network methods of risk assessment of space-time grid is based under a kind of typhoon disaster | |
| Yao | Application of GIS remote sensing information integration in eco-environmental quality monitoring | |
| CN114842163A (en) | Multi-scale dynamic gridding model construction method for disaster assessment | |
| Kaoje | Application of Geographical Information System Techniques in Urban Flood Risk Assessment and Vulnerability Mapping. A Case Study of Cardiff, Wales | |
| Benchelha et al. | Modeling dynamic urban growth using cellular automata and geospatial technique: case of Casablanca in Morocco | |
| Raynor et al. | Multi-hazard analysis and mapping of coastal Tauranga in support of resilience planning | |
| CN118135424B (en) | Ecological environment supervision method and device based on remote sensing image and GIS | |
| Kumar et al. | Cellular automata and Genetic Algorithms based urban growth visualization for appropriate land use policies | |
| Hai et al. | Utilization of Geographic Information System (GIS) for urban management of RachGia City, Vietnam | |
| Nguyen | An Evaluation of Coastal Flooding Risk due to Storm Surge under Future Sea Level Rise Scenarios in Thua Thien Hue Province, Vietnam | |
| Liu et al. | Gully Identification of Debris Flow Disaster Based on Knowledge Distillation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |