CN108304809B - Near real-time earthquake damage assessment method based on post-earthquake aerial image - Google Patents
Near real-time earthquake damage assessment method based on post-earthquake aerial image Download PDFInfo
- Publication number
- CN108304809B CN108304809B CN201810119671.8A CN201810119671A CN108304809B CN 108304809 B CN108304809 B CN 108304809B CN 201810119671 A CN201810119671 A CN 201810119671A CN 108304809 B CN108304809 B CN 108304809B
- Authority
- CN
- China
- Prior art keywords
- building
- earthquake
- simulated
- collapse
- distribution map
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
Abstract
The invention discloses a near real-time earthquake damage assessment method based on post-earthquake aerial images. The method utilizes important information provided by the on-site aerial photo of the earthquake region, improves the precision of earthquake damage evaluation, has extremely high calculation efficiency, and can provide a near-real-time evaluation result within 48 hours after the earthquake.
Description
Technical Field
The invention relates to the technical field of civil engineering, in particular to a near real-time earthquake damage assessment method based on post-earthquake aerial images.
Background
China is a country with multiple earthquakes, and statistical data show that economic losses brought by earthquakes to China are ranked the second worldwide in the year of 1900-2016. The method can quickly and accurately predict the economic loss caused by earthquake building damage, and has great value for making reasonable disaster relief and reconstruction schemes.
The earthquake loss statistical method mainly comprises the steps of on-site investigation or spot check statistical loss, loss prediction according to a loss prediction model, loss prediction according to remote sensing or aerial data and the like. The field survey is relatively the most accurate, but is time-consuming, often weeks or months long, and is susceptible to human factors. The loss prediction model or aerial photo prediction loss is short in time consumption, and the real requirement for quickly evaluating earthquake loss after an earthquake is possibly met. But the loss prediction model requires reasonable seismic motion input; aerial photographs are difficult to identify the damage condition inside the building, and underestimation loss is caused. Therefore, the existing method is not enough to meet the requirements of efficiency and precision of the earthquake damage assessment.
Disclosure of Invention
The object of the present invention is to solve at least to some extent one of the above mentioned technical problems.
Therefore, a first objective of the present invention is to provide a near real-time earthquake damage assessment method based on post-earthquake aerial images, which includes identifying an actual building collapse distribution map of an earthquake area from the post-earthquake aerial images, acquiring a large number of simulated building collapse distribution maps through earthquake damage simulation, selecting an optimal simulated building collapse distribution map most matched with the actual building collapse distribution map from the large number of simulated building collapse distribution maps through a similarity matching method, and performing earthquake damage assessment according to the optimal simulated building collapse distribution map. The method utilizes important information provided by the on-site aerial photo of the earthquake region, improves the precision of earthquake damage evaluation, has extremely high calculation efficiency, and can provide a near-real-time evaluation result within 48 hours after the earthquake.
In order to achieve the above object, a near real-time earthquake damage assessment method based on post-earthquake aerial images according to an embodiment of the first aspect of the present invention includes:
identifying the actual building collapse state of each building in the aerial image after earthquake, and acquiring an actual building collapse distribution map of the earthquake area, wherein the actual building collapse distribution map comprises the actual building collapse state of each building in the earthquake area;
performing earthquake damage simulation on the building model in the earthquake region according to at least one nonlinear time history analysis working condition to obtain each simulated building collapse distribution map of the earthquake region, wherein one nonlinear time history analysis working condition corresponds to one simulated building collapse distribution map, and each simulated building collapse distribution map comprises a simulated detailed damage state of each building in the earthquake region;
matching each simulated building collapse distribution map with an actual building collapse distribution map to obtain a similarity score of each simulated building collapse distribution map;
selecting an optimal simulation building collapse distribution map according to the similarity score of each simulation building collapse distribution map;
and performing earthquake damage evaluation on each building according to the simulated detailed damage state of each building in the optimal simulated building collapse distribution map.
The method for evaluating the earthquake damage of each building according to the simulated detailed damage state of each building in the optimal simulated building collapse distribution map comprises the following steps:
for each building:
according to the formula Lh＝S×DhX P calculating house damage loss LhWherein S is the building area, P is the house replacement unit price, DhSimulating the house damage loss ratio corresponding to the detailed damage state for the optimal;
according to the formula Ld＝γ1×γ2×(ξ×S)×DdX (η XP) calculating the fitment failure loss LdWherein γ is1Correction factor, gamma, to take account of differences in economic conditions in various regions2For considering the correction coefficients of different building applications, ξ is the ratio of the building area of the medium-high grade decorated house to the total house, η is the ratio of the house decoration cost to the house main body cost, DdSimulating the decoration damage loss ratio corresponding to the detailed damage state for the optimal;
according to house damage loss LhAnd loss of finishing damage LdAnd calculating the earthquake economic loss of the house of each building.
In the above method, the matching of each simulated building collapse distribution map with the actual building collapse distribution map to obtain the similarity score of each simulated building collapse distribution map includes:
for the ith simulated building collapse distribution map:
according to the formula
Calculating the similarity score S of the ith simulation building collapse distribution map corresponding to the jth buildingaij；
According to the formula
Calculating the similarity score S of the ith simulated building collapse distribution diagramAi；
Wherein, yijSimulated detailed failure states, y, for the ith building corresponding to the ith simulated building collapse profilejAnd (3) setting the actual building collapse state of the jth building, wherein i is a positive integer, m is the total number of buildings in the earthquake region, j is a positive integer, and j takes a value from 1 to m.
In the above method, the matching of each simulated building collapse distribution map with the actual building collapse distribution map to obtain the similarity score of each simulated building collapse distribution map includes:
for the ith simulated building collapse distribution map:
according to the formula
Calculating the similarity score S of the ith simulation building collapse distribution map corresponding to the jth buildingbij；
According to the formula
Calculating the similarity score S of the ith simulated building collapse distribution diagramBi；
Wherein the weight of the jth building
pjThe probability of collapse for the jth building;
wherein, yijSimulated detailed failure states, y, for the ith building corresponding to the ith simulated building collapse profilejAnd (3) setting the actual building collapse state of the jth building, wherein i is a positive integer, m is the total number of buildings in the earthquake region, j is a positive integer, and j takes a value from 1 to m.
The method as described above, further comprising:
according to the formula
Calculating the collapse probability p of the jth buildingj；
Wherein x is a collapse probability factor vector, theta is a parameter vector theta obtained based on logic classification algorithm training in machine learning, and the value range of a logic function h (z) is (0, 1).
The method as described above, further comprising:
determining a building collapse distribution model P (y | x; theta) ═ h (theta)Tx)y[1-h(θTx)]1-y；
The maximum likelihood estimation value model of the parameter vector theta is determined according to the building collapse distribution model as
Selecting a training set from the sample data set, training a maximum likelihood estimation value model by using the training set to obtain a maximum likelihood estimation value of a parameter vector theta, wherein a collapse probability factor x of the jth buildingjAnd the actual building collapsed state yjJ sample data as a sample data set;
and determining the parameter vector theta corresponding to the maximum likelihood estimation value as the parameter vector theta obtained based on the logic classification algorithm training in machine learning.
The method for selecting the optimal simulated building collapse distribution map according to the similarity scores of the simulated building collapse distribution maps comprises the following steps:
and sequencing the similarity scores of the simulated building collapse distribution maps, and selecting the simulated building collapse distribution map meeting the preset conditions as the optimal simulated building collapse distribution map.
The method for performing earthquake damage simulation on the building model in the earthquake region according to at least one nonlinear time history analysis working condition to obtain each simulated building collapse distribution map of the earthquake region comprises the following steps:
determining a building model of each building in the earthquake region, wherein the building model is a multi-mass-point shearing series model for medium and low-rise buildings in the earthquake region, and the building model is a multi-mass-point parallel shearing bending coordination model for high-rise buildings in the earthquake region;
analyzing the working condition for each nonlinear time history: carrying out earthquake damage simulation on the corresponding building model by using the amplitude-modulated earthquake motion corresponding to each building to obtain a simulated detailed damage state of each building;
and outputting a simulated building collapse distribution diagram corresponding to each nonlinear time history analysis working condition according to the simulated detailed damage state of each building.
The method as described above, further comprising:
determining at least one ground motion prediction equation;
acquiring at least one earthquake motion record from an earthquake motion database;
amplitude modulating the acquired at least one seismic motion recording by using the determined at least one ground motion prediction equation to construct at least one nonlinear time history analysis condition.
According to the method, the image processing is carried out on the post-earthquake aerial image to obtain the actual building collapse distribution map of the earthquake area, and the method comprises the following steps:
and identifying the actual building collapse state of each building in the aerial image after the earthquake by using a visual interpretation method, and obtaining the actual building collapse distribution map of the earthquake area according to the building distribution map of the earthquake area and the identified actual building collapse state of each building.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The above and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which,
FIG. 1 is a schematic flow chart illustrating a near real-time damage assessment method based on post-earthquake aerial images according to an embodiment of the present invention;
FIG. 2 is a schematic illustration of an exemplary collapse profile;
FIG. 3 is a plot of actual building collapse distribution for a blund earthquake;
FIG. 4 is a plot of optimal simulated building collapse profiles for the meadow earthquake;
FIG. 5 is a graph of similarity scores versus economic loss;
fig. 6 is a bar graph of economic losses.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.
The near real-time earthquake damage assessment method based on post-earthquake aerial images according to the embodiment of the invention is described below with reference to the accompanying drawings.
Fig. 1 is a schematic flow chart of a near real-time earthquake damage assessment method based on post-earthquake aerial images according to an embodiment of the present invention. As shown in fig. 1, the near real-time earthquake damage assessment method based on post-earthquake aerial images provided in this embodiment includes the following steps:
s101, identifying actual building collapse states of all buildings in the aerial images after the earthquake, and obtaining an actual building collapse distribution map of the earthquake area, wherein the actual building collapse distribution map comprises the actual building collapse states of all the buildings in the earthquake area.
Specifically, the unmanned aerial vehicle can be used for aerial photography of the earthquake area to obtain the post-earthquake aerial photography image, and then the post-earthquake aerial photography image is subjected to image preprocessing such as enhancement, restoration and compression to obtain a clear and visible post-earthquake aerial photography image. In the embodiment, the actual building collapse states include two states, the first state is that the building collapses and the building does not collapse, and the actual building collapse distribution map of the earthquake area is output based on the identified actual building collapse states of the buildings.
Specifically, the visual interpretation method is that people extract and analyze the comprehensive information of a detection target in an image through comprehensive analysis, logical reasoning, verification and detection by applying abundant professional background knowledge and observing through naked eyes. According to the method, the actual building collapse states of all buildings are more accurately identified from the aerial images after the earthquake through strict customs control of professionals, and then the actual building collapse distribution map of the earthquake area is determined by combining the building distribution map of the earthquake area.
S102, performing earthquake damage simulation on the building model in the earthquake region according to at least one nonlinear time history analysis working condition to obtain each simulated building collapse distribution map of the earthquake region, wherein one nonlinear time history analysis working condition corresponds to one simulated building collapse distribution map, and each simulated building collapse distribution map comprises a simulated detailed damage state of each building in the earthquake region.
Specifically, a simulated building collapse distribution map of the earthquake area is obtained through nonlinear time history simulation, and the simulated detailed damage state of each building is reflected in a simulation result, namely the simulated building collapse distribution map. In this embodiment, simulating detailed failure states may include sound, mild, moderate, severe, collapsed, etc.
Before the earthquake damage simulation of the buildings in the earthquake areas is executed, massive nonlinear time history analysis working conditions need to be constructed.
How to construct the nonlinear time history analysis condition is described below.
In a possible implementation manner, the specific implementation manner of "constructing the nonlinear time history analysis condition" is: determining at least one ground motion prediction equation; acquiring at least one earthquake motion record from an earthquake motion database; amplitude modulating the acquired at least one seismic motion recording by using the determined at least one ground motion prediction equation to construct at least one nonlinear time history analysis condition.
For example, p (e.g., p is a positive integer greater than 30) different Ground Motion Prediction Equations (GMPEs) are selected, and then q (q is a positive integer greater than 20) seismic Motion records are selected from a seismic Motion database for each GMPE according to the seismic source parameters, so that pq nonlinear time history analysis conditions can be constructed.
In this embodiment, the specific implementation manner of step S102 is:
and S21, determining building models of all buildings in the earthquake area.
Specifically, when building a building model, a multi-mass-point shearing series model is adopted to simulate middle and low-rise buildings in an earthquake area, and a multi-mass-point parallel shearing bending coordination model is adopted to simulate high-rise buildings in the earthquake area.
S22, analyzing the working condition of each nonlinear time history: and performing earthquake damage simulation on the corresponding building model by using the amplitude-modulated earthquake motion corresponding to each building as input to obtain the simulated detailed damage state of each building.
Specifically, a plurality of nonlinear time history analysis working conditions are constructed in advance, specifically, the earthquake motion intensity of the position of each building is calculated according to a ground motion prediction equation, amplitude modulation is carried out on each selected earthquake motion record according to the calculated earthquake motion intensity, and therefore the plurality of nonlinear time history analysis working conditions are constructed.
For example, the total number of the nonlinear time history analysis working conditions is 60, and the total number of buildings in the earthquake region is 50. Analyzing the working condition of the 1 st nonlinear time history, wherein the result of the earthquake damage simulation is as follows: building 1: intact, building 2: light, building 3: medium, architecture 4: severe, architecture 5: collapse, building 6: collapse, building 7: collapse, building 8: severe … … building 49: severe, building 50: and (6) collapsing. And analogizing in turn to obtain the earthquake damage simulation results from the 2 nd nonlinear time history analysis working condition to the 60 th nonlinear time history analysis working condition.
And S23, outputting a simulated building collapse distribution diagram corresponding to each nonlinear time history analysis working condition according to the simulated detailed damage state of each building.
For example, the 1 st nonlinear time history analysis condition respectively obtains the simulated detailed failure states of the buildings 1 to 50, and the simulated detailed failure states of the buildings 1 to 50 are reflected in the simulated building collapse distribution map. And analogizing in turn to obtain the simulated building collapse distribution map from the 2 nd nonlinear time history analysis working condition to the 60 th nonlinear time history analysis working condition.
In the embodiment, the earthquake damage simulation is carried out by analyzing the working condition by utilizing the mass nonlinear time history, so that a plurality of simulated building collapse distribution maps are obtained, and a foundation is provided for the similar analysis of the building collapse distribution.
S103, matching each simulated building collapse distribution map with the actual building collapse distribution map to obtain the similarity score of each simulated building collapse distribution map.
In this embodiment, whether the building collapses is a binary event, so the similarity between the simulated building collapse distribution map and the actual building collapse distribution map can be measured by using binary similarity. In order to find the distribution map which is most matched with the actual building collapse distribution map from a group of simulated building collapse distribution maps, the implementation calculates the similarity score of the simulated building collapse distribution maps, and finds the distribution map which is most matched with the actual building collapse distribution map according to the height of the similarity score.
Specifically, a point-by-point matching method can be adopted to calculate the similarity score of the simulated building collapse distribution map, and the method is simple and effective. In the point-by-point matching method, whether the simulated detailed destruction state of each building is the same as the actual building collapse state or not is compared one by one, if so, 1 score is counted, and otherwise, 0 score is counted.
Let the random variable y represent the building collapse situation, and y follows the Bernoulli distribution, i.e., y-B (1, p). y-1 indicates collapse, and y-0 indicates no collapse.
Assuming that the total number of buildings in the earthquake area is m, wherein m is a positive integer, the total number of the obtained simulated building collapse distribution graphs is I, and I is a positive integer.
For the jth building, yijSimulated detailed failure states, y, for the ith building corresponding to the ith simulated building collapse profilejAnd the actual building collapse state of the jth building is shown, wherein I is a positive integer, values are obtained from 1 to I, j is a positive integer, and values of j are obtained from 1 to m.
Then, for the point-by-point matching method, the specific implementation manner of step S103 is:
s31, for the ith simulation building collapse distribution diagram:
according to the formula:
calculating the similarity score S of the ith simulation building collapse distribution map corresponding to the jth buildingaij。
Specifically, for any one simulated building collapse distribution map, whether the simulated detailed damage state of each building is the same as the actual building collapse state or not is judged, if yes, the similarity score between the simulated detailed damage state of the building and the actual building collapse state is 1 score, and otherwise, the similarity score is 0 score.
S32, according to the formula:
calculating the similarity score S of the ith simulated building collapse distribution diagramAi。
Specifically, for any of the simulated building collapse distribution maps, according to step S31, the similarity scores of the simulated detailed damage states and the actual building collapse states of the buildings can be obtained, and the similarity scores of the simulated detailed damage states and the actual building collapse states of the buildings are summed and then averaged to obtain the similarity score of the simulated building collapse distribution map.
Specifically, the embodiment may also use a weighted point-by-point matching method to calculate the similarity score of the simulated building collapse distribution map.
The probability of collapse is different for different buildings, depending on the building characteristics and building location. In order to introduce the influence of other important information such as the position coordinates, the structure types and the like of the buildings on the result, a weight related to the collapse probability of the buildings is defined, and the similarity score of the simulated detailed damage state and the actual collapse state of each building is multiplied by the weight to form the final similarity score of the simulated detailed damage state and the actual collapse state of each building.
Let the random variable y represent the building collapse state, y obeys the Bernoulli distribution, i.e., y-B (1, p). y-1 indicates collapse, and y-0 indicates no collapse.
Assuming that the total number of buildings in the earthquake area is m, wherein m is a positive integer, the total number of the obtained simulated building collapse distribution graphs is I, and I is a positive integer.
For the jth building, yijSimulated detailed failure states, y, for the ith building corresponding to the ith simulated building collapse profilejActual building collapse state for jth buildingWherein I is a positive integer, and takes a value from 1 to I, j is a positive integer, and j takes a value from 1 to m.
Then, the specific implementation manner of step S103 is:
s33, for the ith simulation building collapse distribution diagram:
according to the formula:
calculating the similarity score S of the ith simulation building collapse distribution map corresponding to the jth buildingbij。
Specifically, for any one simulated building collapse distribution map, whether the simulated detailed damage state of each building is the same as the actual building collapse state or not is judged, if yes, the similarity score between the simulated detailed damage state of the building and the actual building collapse state is 1 score, and otherwise, the similarity score is 0 score.
S34, according to the formula:
calculating the similarity score S of the ith simulated building collapse distribution diagramBi。
Wherein the weight of the jth building
pjThe collapse probability of the jth building.
Fig. 2 is a schematic diagram of an exemplary collapse profile. In fig. 2, 12 buildings are included. Fig. 2(a) shows an actual collapse distribution, fig. 2(b) shows a simulated collapse distribution 1, fig. 2(c) shows a simulated collapse distribution 2, and fig. 2(d) shows a simulated collapse distribution 3.
When the similarity score is calculated by adopting a point-by-point matching method, the difference between the simulated collapse distribution 1 and the actual collapse distribution is the largest, and the corresponding similarity score is lower (S)A17/12 points), a similarity score (S) corresponding to the simulated collapse distribution 2 is obtainedA210/12 points) is high, simulating collapse distribution 3Corresponding similarity score (S)A310/12 points) higher.
When the similarity score is calculated by adopting a weighted point-by-point matching method, the similarity score S corresponding to the collapse distribution 1 is simulatedB1Similarity score S corresponding to simulated collapse distribution 2 of 0.596 pointsB2Similarity score S corresponding to simulated collapse distribution 3 of 0.857 pointB3Score 0.882. It can be found that the weighted point-by-point matching method can take into account the influence of other important information such as the position coordinates, the structure type and the like of the building on the result relative to the point-by-point matching method.
In the present embodiment, the collapse probability of each building can be obtained based on a machine learning method.
First, the determining factor for determining the building collapse probability is the collapse probability factor vector x. For example, if it is determined that the factors such as building coordinates, structure type, construction time, number of floors, etc. are the determining factors for building collapse, the collapse probability factor vector x is constructed by the building structure, structure type, construction time, number of floors, but is not limited to the illustration.
Next, a building collapse probability is determined. Specifically, the method comprises the following steps:
probability of collapse p for jth buildingjComprises the following steps:
then, the non-collapse probability 1-p of the jth buildingjComprises the following steps:
1-pj＝P(y＝0|x＝xj；θ)＝1-h(θTxj)(6)；
wherein x is a collapse probability factor vector, xjAnd representing a collapse probability factor vector of the jth building, wherein theta is a parameter vector theta obtained based on the training of a logic classification algorithm, and the value range of a logic function h (z) is (0, 1).
As can be seen from the above equation, the x collapse probability factor vector is a predetermined determining factor and is a known quantity, and as long as the parameter vector θ is determined, the collapse probability and the non-collapse probability of each building can be obtained according to the equation (5) and the equation (6).
The following describes how to train the obtained parameter vector θ based on the logical classification algorithm in machine learning.
In one possible implementation manner, the specific implementation manner of the "parameter vector θ trained based on the logic classification algorithm in machine learning" is as follows:
s201, determining a building collapse distribution model:
P(y|x；θ)＝h(θTx)y[1-h(θTx)]1-y(7)。
specifically, through the observation of the formula (5) and the formula (6), the building collapse distribution model is obtained by combining the two.
S202, determining a maximum likelihood estimation value model of the parameter vector theta according to the building collapse distribution model as follows:
specifically, according to the specification of the likelihood function in the related art, after the building collapse distribution model is determined, it is not difficult to obtain the maximum likelihood estimation value model of the parameter vector θ.
S203, selecting a training set from the sample data set, and training the maximum likelihood estimation value model by using the training set to obtain the maximum likelihood estimation value of the parameter vector theta.
Specifically, the total number of buildings in the earthquake area is m, the actual building collapse states of all the buildings are identified from the post-earthquake aerial images, and collapse probability factors of all the buildings are determined. For example, for the jth building, the collapse probability factor is xjThe actual building collapse state is yj. According to known information, a sample data set with the total number of samples being m is constructed, and the collapse probability factor x of the jth building is determinedjAnd the actual building collapsed state yjThe jth sample data as the sample data set.
In this embodiment, a part of samples from the sample data set may be selected as a training set, and the maximum likelihood estimation model may be trained to obtain the maximum likelihood estimation of the parameter vector θ
Further, to avoid the overfitting phenomenon, a non-negative regularization parameter λ is added in equation (8), forming equation (9):
in this embodiment, a part of samples may be selected from the sample data set as a cross validation set, and a value of the regularization parameter λ may be determined. Of course, a part of samples can be selected from the sample data set to serve as a cross validation set test set, and the machine learning precision can be tested.
And S204, determining the parameter vector theta corresponding to the maximum likelihood estimation value as the parameter vector theta obtained based on the logic classification algorithm training in machine learning.
Specifically, after the parameter vector θ is obtained, the collapse probability and non-collapse probability of each building can be obtained by substituting the parameter vector θ into the formula (5) and the formula (6).
In the embodiment, the collapse probability and the non-collapse probability of each building are calculated by using the parameter vector theta obtained by training based on the logic classification algorithm in machine learning, and the calculation result is more reliable.
And S104, selecting the optimal simulated building collapse distribution map according to the similarity score of each simulated building collapse distribution map.
Specifically, the similarity score of each simulated building collapse distribution map may be compared with a preset threshold, and the simulated building collapse distribution map with the similarity score larger than the preset threshold may be used as the optimal simulated building collapse distribution map, or the similarity scores of each simulated building collapse distribution map may be sorted, and the simulated building collapse distribution map with the highest similarity score may be used as the optimal simulated building collapse distribution map, but the invention is not limited thereto.
In a possible implementation manner, the specific implementation manner of step S104 is: and sequencing the similarity scores of the simulated building collapse distribution maps, and selecting the simulated building collapse distribution map meeting the preset conditions as the optimal simulated building collapse distribution map. In this embodiment, the preset conditions may be set by themselves, for example, the similarity scores of the simulated building collapse distribution maps are sorted in a descending order, and the simulated building collapse distribution maps with a preset number of similarity scores arranged in front are used as the optimal simulated building collapse distribution map (that is, there are a plurality of optimal simulated building collapse distribution maps); for example, the simulated building collapse distribution map with the highest similarity score is used as the optimal simulated building collapse distribution map; and if a plurality of simulation building collapse distribution graphs with the highest similarity scores exist, taking the simulation building collapse distribution graphs with the highest similarity scores as the optimal simulation building collapse distribution graph.
And S105, performing earthquake damage evaluation on each building according to the simulated detailed damage state of each building in the optimal simulated building collapse distribution map.
In particular, the house earthquake economic loss is house damage loss LhAnd loss of finishing damage LdAnd the method for calculating the house earthquake economic loss is formulated according to the actual situation.
In this embodiment, for each building, according to the formula:
Lh＝S×Dh×P(10)
calculating house damage loss Lh。
In the above formula, S is the building area in m2(ii) a P is house reset unit price, unit is Yuan/m2，DhAnd optimally simulating the house damage loss ratio corresponding to the detailed damage state.
Table 1 shows the median house failure and fitment failure loss ratios for each failure condition. In this embodiment, DhValues were taken according to table 1.
TABLE 1 median value of house destruction and decoration destruction loss ratio in each destruction state
Table 2 shows the median construction replacement cost. In this embodiment, P is taken according to table 2.
Table 2 median construction reset cost, unit: yuan/m2
In this embodiment, for each building, according to the formula:
Ld＝γ1×γ2×(ξ×S)×Dd×(η×P)(11)
calculating decoration damage loss Ld。
In the above formula, γ1Correction factor, gamma, to take account of differences in economic conditions in various regions2In order to consider correction coefficients of different building applications, if not, 1.0 is taken, ξ is the proportion of the building area of the medium-high-grade decorated house to the total house, η is the ratio of the house decoration cost to the house main body cost, DdAnd optimally simulating the decoration damage loss ratio corresponding to the detailed damage state. In this embodiment, DdValues were taken according to table 1.
It should be noted that, as can be seen from table 1, the building economic loss calculation requires a detailed destruction state of the building as an input, and the collapse distribution of the building is only identified from the aerial photograph of the disaster area and is not sufficient for calculating the economic loss. The collapse distribution of the building can be used to identify the optimal simulation result from a large number of simulation results, which contains the detailed destruction state of the building, and thus can be used to calculate the economic loss.
According to the near real-time earthquake damage assessment method based on the post-earthquake aerial images, the actual building collapse distribution map of the earthquake area is identified from the post-earthquake aerial images, a large number of simulated building collapse distribution maps are obtained through earthquake damage simulation, then the optimal simulated building collapse distribution map which is most matched with the actual building collapse distribution map is selected from the large number of simulated building collapse distribution maps through a similarity matching method, and earthquake damage assessment is carried out according to the optimal simulated building collapse distribution map. The method utilizes important information provided by the on-site aerial photo of the earthquake region, improves the precision of earthquake damage evaluation, has extremely high calculation efficiency, and can provide a near-real-time evaluation result within 48 hours after the earthquake.
The near real-time earthquake damage assessment method based on post-earthquake aerial images is further described below by taking the meadow earthquake as an example.
FIG. 3 is a plot of actual building collapse distribution for a blund earthquake; FIG. 4 is a plot of optimal simulated building collapse profiles for the meadow earthquake; FIG. 5 is a graph of similarity scores versus economic loss; fig. 6 is a bar graph of economic losses.
Taking the construction of the Bingshan town of the Ludian earthquake in 2014 as an example, the earthquake source information of the Ludian earthquake can be obtained by inquiring related documents: the seismic magnitude is M6.5, the seismic source depth is 12km, the epicenter is 27.189 degrees N and 103.409 degrees E, and the earthquake causes serious damage to the dragon mountain town beyond 9 km. In the next day after the earthquake, the Chinese earthquake administration acquires a large number of aerial photos of the disaster area by using the unmanned aerial vehicle; the detailed damage information of the building is obtained by a specialist going to the site to carry out earthquake damage investigation for nearly one month.
The process of applying the near real-time earthquake damage assessment method based on the post-earthquake aerial images to the meadow earthquake is described below.
First, the actual building collapse profile of the meadow earthquake is acquired (i.e., fig. 3). Specifically, the actual building collapse distribution map of the Bingwan town of Ludian (namely, FIG. 3) can be obtained quickly by identifying a large number of disaster area aerial photographs released by the China earthquake administration and on-site aerial photographs provided by various big news media.
Next, a linear attenuation function (12) as the formula is defined, and taking the coordinate system shown in fig. 3 as an example, in the coordinate system shown in fig. 3, the positive direction of the x ' axis is east, the positive direction of the y ' axis is north (N), and the maximum value of the coordinates in the y ' axis direction is y ' with respect to the plane coordinates (y ', x ') of the building center point 'maxThe maximum value of the coordinates in the x 'axis direction is x'maxThe minimum value of the coordinate in the y 'axis direction is y'minThe minimum value of the coordinates in the x 'axis direction is x'min。
PGAmaxIs a coefficient whose value range is {0.2g,0.4g,0.6g,0.8g,1.0g,1.2g }, thus defining 30 different ground transportsAnd (4) a dynamic prediction equation. And selecting 28 near-field seismic records from a seismic database, thereby constructing 840 different nonlinear time history analysis working conditions.
A multi-quality-point shearing series model is adopted to simulate middle and low-rise buildings, a multi-quality-point parallel shearing bending coordination model is adopted to simulate high-rise buildings, 840 different nonlinear time history analysis working conditions are adopted to simulate nonlinear time histories of earthquake damage of regional buildings, and detailed damage states (intact, slight, medium, serious and collapsed) of each building are obtained.
Then, using the collapse distribution similarity matching method, the score of each simulated building collapse distribution map and the optimal simulated building collapse distribution map, that is, fig. 4, can be calculated.
Finally, the similarity score is plotted against the economic loss, as shown in fig. 5. And the statistics of FIG. 5 are performed to form a histogram of economic losses as in FIG. 6, V in FIG. 6optAnd calculating to obtain economic loss according to the simulated detailed damage state of each building in the optimal simulated building collapse distribution diagram.
In FIG. 5, except for the simulated loss VjAlso adding the actual loss VactualVulnerability matrix loss VyinLoss of actual seismic behavior Vrecorded。
It should be noted that the simulated loss VjAnd calculating to obtain economic loss according to the simulated detailed damage state of each building in the simulated building collapse distribution diagram.
It should be noted that the actual loss VactualThe earthquake damage investigation is carried out for experts on site and the economic loss is obtained after nearly one month.
It should be noted that the Chinese strong vibration table net obtains the actual earthquake motion near the field, and the nonlinear time history analysis is performed by using the actual earthquake motion recorded by the Chinese strong vibration table net to obtain the actual earthquake condition loss Vrecorded。
To be pointed outThe building loss of the earthquake dragon-headed mountain town is evaluated by adopting a traditional vulnerability matrix method, and the economic loss is recorded as vulnerability matrix loss Vyin。
From the results it can be seen that:
(1) the optimal simulated building collapse distribution map (namely, figure 4) obtained by the weighted point-by-point matching method provided by the invention is still similar to the actual building collapse distribution map (namely, figure 3).
(2) Loss V corresponding to the optimal simulation result obtained by using the method of the inventionoptAnd actual loss VactualClose to and far better than the loss V given by the vulnerability matrix methodyinAs shown in fig. 6. The key reason is that the method provided by the invention fully utilizes important information of the actual collapse distribution condition of the building.
(3) Loss prediction result V obtained using measured blund seismic recordsrecordedAnd actual loss VactualThe fit was also good. It should be noted that the seismic record actually measured by the strong earthquake station may not fully reflect the actual earthquake action on each building, and therefore may not be completely consistent with the actual earthquake damage.
It should be noted that, the 840 nonlinear time history analysis conditions are operated in parallel on one multi-core computer (CPU: Intel E5-2695v4@2.10Hz, 36 cores; memory: 64GB), the time consumption is only about 4min, and collapse distribution similarity matching can be completed within several seconds. With the improvement of remote sensing or unmanned aerial vehicle technology, a large number of satellites or aerial photographs in disaster areas can be expected to be obtained within 24 hours after the earthquake, so that the collapse condition of the buildings in the disaster areas can be rapidly identified.
In summary, from the above examples, it can be summarized that: the method provided by the invention is a near-real-time earthquake economic loss assessment method, and can give earthquake loss estimation close to the expert survey result which takes weeks within one or two days after an earthquake by fully utilizing important information provided by aerial photos in disaster areas.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "a plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing steps of a custom logic function or process, and alternate implementations are included within the scope of the preferred embodiment of the present invention in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present invention.
The logic and/or steps represented in the flowcharts or otherwise described herein, e.g., an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.
It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be stored in a computer readable storage medium, and the program may be executed by a computer to instruct the relevant hardware, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.
In addition, functional units in the embodiments of the present invention may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a separate product, may also be stored in a computer readable storage medium.
The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc. Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.
Claims (10)
1. A near real-time seismic damage assessment method based on post-earthquake aerial images is characterized by comprising the following steps:
identifying the actual building collapse state of each building in the aerial image after earthquake, and acquiring an actual building collapse distribution map of the earthquake area, wherein the actual building collapse distribution map comprises the actual building collapse state of each building in the earthquake area;
performing earthquake damage simulation on the building model in the earthquake region according to at least one nonlinear time history analysis working condition to obtain each simulated building collapse distribution map of the earthquake region, wherein one nonlinear time history analysis working condition corresponds to one simulated building collapse distribution map, and each simulated building collapse distribution map comprises a simulated detailed damage state of each building in the earthquake region;
matching each simulated building collapse distribution map with an actual building collapse distribution map to obtain a similarity score of each simulated building collapse distribution map;
selecting an optimal simulation building collapse distribution map according to the similarity score of each simulation building collapse distribution map;
performing earthquake damage evaluation on each building according to the simulated detailed damage state of each building in the optimal simulated building collapse distribution map;
wherein, the creation mode of the at least one nonlinear time history analysis working condition is as follows: determining at least one ground motion prediction equation; acquiring at least one earthquake motion record from an earthquake motion database; amplitude modulating the acquired at least one seismic motion recording by using the determined at least one ground motion prediction equation to construct at least one nonlinear time history analysis condition.
2. The method of claim 1, wherein said assessing the damage from earthquake of each building based on the simulated detailed failure status of each building in the optimal simulated building collapse profile comprises:
for each building:
according to the formula Lh＝S×DhX P calculating house damage loss LhWherein S is the building area, P is the house replacement unit price, DhSimulating the house damage loss ratio corresponding to the detailed damage state for the optimal;
according to the formula Ld＝γ1×γ2×(ξ×S)×DdX (η XP) calculating the fitment failure loss LdWherein γ is1Correction factor, gamma, to take account of differences in economic conditions in various regions2For considering the correction coefficients of different building applications, ξ is the ratio of the building area of the medium-high grade decorated house to the total house, η is the ratio of the house decoration cost to the house main body cost, DdSimulating the decoration damage loss ratio corresponding to the detailed damage state for the optimal;
according to house damage loss LhAnd loss of finishing damage LdAnd calculating the earthquake economic loss of the house of each building.
3. The method of claim 1, wherein matching each simulated building collapse profile to an actual building collapse profile to obtain a similarity score for each simulated building collapse profile comprises:
for the ith simulated building collapse distribution map:
according to the formula
Calculating the similarity score S of the ith simulation building collapse distribution map corresponding to the jth buildingaij；
According to the formula
Calculating the similarity score S of the ith simulated building collapse distribution diagramAi；
Wherein, yijSimulated detailed failure states, y, for the ith building corresponding to the ith simulated building collapse profilejAnd (3) setting the actual building collapse state of the jth building, wherein i is a positive integer, m is the total number of buildings in the earthquake region, j is a positive integer, and j takes a value from 1 to m.
4. The method of claim 1, wherein matching each simulated building collapse profile to an actual building collapse profile to obtain a similarity score for each simulated building collapse profile comprises:
for the ith simulated building collapse distribution map:
according to the formula
Calculating the similarity score S of the ith simulation building collapse distribution map corresponding to the jth buildingbij；
According to the formula
Calculating the similarity score S of the ith simulated building collapse distribution diagramBi；
Wherein the weight of the jth building
pjThe probability of collapse for the jth building;
wherein, yijModel for simulating building collapse distribution diagram of ith building corresponding to jth buildingQuasi detailed destruction state, yjAnd (3) setting the actual building collapse state of the jth building, wherein i is a positive integer, m is the total number of buildings in the earthquake region, j is a positive integer, and j takes a value from 1 to m.
5. The method of claim 4, further comprising:
according to the formula
Calculating the collapse probability p of the jth buildingj；
Wherein x is a collapse probability factor vector, theta is a parameter vector theta obtained based on logic classification algorithm training in machine learning, and the value range of a logic function h (z) is (0, 1).
6. The method of claim 4, further comprising:
determining a building collapse distribution model P (y | x; theta) ═ h (theta)Tx)y[1-h(θTx)]1-y；
The maximum likelihood estimation value model of the parameter vector theta is determined according to the building collapse distribution model as
Selecting a training set from the sample data set, training a maximum likelihood estimation value model by using the training set to obtain a maximum likelihood estimation value of a parameter vector theta, wherein a collapse probability factor x of the jth buildingjAnd the actual building collapsed state yjJ sample data as a sample data set;
and determining the parameter vector theta corresponding to the maximum likelihood estimation value as the parameter vector theta obtained based on the logic classification algorithm training in machine learning.
7. The method of claim 1, wherein selecting the optimal simulated building collapse profile based on the similarity scores of the respective simulated building collapse profiles comprises:
and sequencing the similarity scores of the simulated building collapse distribution maps, and selecting the simulated building collapse distribution map meeting the preset conditions as the optimal simulated building collapse distribution map.
8. The method of claim 1, wherein said performing a seismic damage simulation of the structure in the seismic region on the building model based on at least one non-linear time history analysis condition to obtain respective simulated building collapse profiles for the seismic region comprises:
determining a building model of each building in the earthquake region, wherein the building model is a multi-mass-point shearing series model for medium and low-rise buildings in the earthquake region, and the building model is a multi-mass-point parallel shearing bending coordination model for high-rise buildings in the earthquake region;
analyzing the working condition for each nonlinear time history: performing earthquake damage simulation on the corresponding building model by using the amplitude-modulated earthquake motion corresponding to each building as input to obtain a simulated detailed damage state of each building;
and outputting a simulated building collapse distribution diagram corresponding to each nonlinear time history analysis working condition according to the simulated detailed damage state of each building.
9. The method of claim 1, further comprising:
determining at least one ground motion prediction equation;
acquiring at least one earthquake motion record from an earthquake motion database;
amplitude modulating the acquired at least one seismic motion recording by using the determined at least one ground motion prediction equation to construct at least one nonlinear time history analysis condition.
10. The method of claim 1, wherein the image processing of the post-earthquake aerial images to obtain the actual building collapse distribution map of the earthquake area comprises:
and identifying the actual building collapse state of each building in the aerial image after the earthquake by using a visual interpretation method, and obtaining the actual building collapse distribution map of the earthquake area according to the building distribution map of the earthquake area and the identified actual building collapse state of each building.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810119671.8A CN108304809B (en) | 2018-02-06 | 2018-02-06 | Near real-time earthquake damage assessment method based on post-earthquake aerial image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810119671.8A CN108304809B (en) | 2018-02-06 | 2018-02-06 | Near real-time earthquake damage assessment method based on post-earthquake aerial image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108304809A CN108304809A (en) | 2018-07-20 |
| CN108304809B true CN108304809B (en) | 2020-03-27 |
Family
ID=62864450
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810119671.8A Active CN108304809B (en) | 2018-02-06 | 2018-02-06 | Near real-time earthquake damage assessment method based on post-earthquake aerial image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108304809B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111985355B (en) * | 2020-08-01 | 2022-09-27 | 桂林理工大学 | Remote sensing building earthquake damage assessment method and system based on deep learning and cloud computing |
| CN114862687B (en) * | 2021-02-04 | 2023-05-09 | 四川大学 | Self-adaptive compressed image restoration method driven by depth deblocking operator |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101979961A (en) * | 2010-05-18 | 2011-02-23 | 中国地震局地球物理研究所 | Disaster condition acquisition system |
| CN201910818U (en) * | 2011-01-05 | 2011-07-27 | 贾群林 | Seismic disaster relief simulation system |
| CN103116810A (en) * | 2013-01-28 | 2013-05-22 | 北京师范大学 | Method for evaluating population loss in earthquake |
| CN103678937A (en) * | 2013-12-29 | 2014-03-26 | 中国地震局工程力学研究所 | Method for evaluating overall earthquake damage level of reinforced concrete frame structure based on equivalent single-degree-of-freedom system |
| CN103761740A (en) * | 2014-01-23 | 2014-04-30 | 武汉大学 | Construction damage assessment method based on single post-earthquake POLSAR image |
| CN104239885A (en) * | 2014-09-05 | 2014-12-24 | 北京航天控制仪器研究所 | Earthquake disaster damage degree evaluation method based on unmanned aerial vehicle aerial photos |
| CN104346538A (en) * | 2014-11-26 | 2015-02-11 | 中国测绘科学研究院 | Earthquake hazard evaluation method based on control of three disaster factors |
| CN105809266A (en) * | 2014-12-31 | 2016-07-27 | 郑静晨 | Method for preliminary evaluation and prediction on earthquake casualty |
| CN106251032A (en) * | 2016-06-28 | 2016-12-21 | 中国人民解放军后勤工程学院 | A kind of Earthquake-landslide disaster Disaster rapid evaluation method |
| CN106651039A (en) * | 2016-12-27 | 2017-05-10 | 中国地震局工程力学研究所 | City building group seismic hazard matrix dynamic prediction method |
| CN106875096A (en) * | 2017-01-17 | 2017-06-20 | 山西省地震局 | A kind of earthquake house loss appraisal system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10909647B2 (en) * | 2015-12-09 | 2021-02-02 | One Concern, Inc. | Damage data propagation in predictor of structural damage |
-
2018
- 2018-02-06 CN CN201810119671.8A patent/CN108304809B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101979961A (en) * | 2010-05-18 | 2011-02-23 | 中国地震局地球物理研究所 | Disaster condition acquisition system |
| CN201910818U (en) * | 2011-01-05 | 2011-07-27 | 贾群林 | Seismic disaster relief simulation system |
| CN103116810A (en) * | 2013-01-28 | 2013-05-22 | 北京师范大学 | Method for evaluating population loss in earthquake |
| CN103678937A (en) * | 2013-12-29 | 2014-03-26 | 中国地震局工程力学研究所 | Method for evaluating overall earthquake damage level of reinforced concrete frame structure based on equivalent single-degree-of-freedom system |
| CN103761740A (en) * | 2014-01-23 | 2014-04-30 | 武汉大学 | Construction damage assessment method based on single post-earthquake POLSAR image |
| CN104239885A (en) * | 2014-09-05 | 2014-12-24 | 北京航天控制仪器研究所 | Earthquake disaster damage degree evaluation method based on unmanned aerial vehicle aerial photos |
| CN104346538A (en) * | 2014-11-26 | 2015-02-11 | 中国测绘科学研究院 | Earthquake hazard evaluation method based on control of three disaster factors |
| CN105809266A (en) * | 2014-12-31 | 2016-07-27 | 郑静晨 | Method for preliminary evaluation and prediction on earthquake casualty |
| CN106251032A (en) * | 2016-06-28 | 2016-12-21 | 中国人民解放军后勤工程学院 | A kind of Earthquake-landslide disaster Disaster rapid evaluation method |
| CN106651039A (en) * | 2016-12-27 | 2017-05-10 | 中国地震局工程力学研究所 | City building group seismic hazard matrix dynamic prediction method |
| CN106875096A (en) * | 2017-01-17 | 2017-06-20 | 山西省地震局 | A kind of earthquake house loss appraisal system |
Non-Patent Citations (1)
| Title |
|---|
| 面向对象的高分辨率遥感影像建筑物震害信息提取;甘甜等;《测绘工程》;20150430;第24卷(第4期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108304809A (en) | 2018-07-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7352632B2 (en) | machine learning tools | |
| CN113496104B (en) | Precipitation prediction correction method and system based on deep learning | |
| Cao et al. | Facade geometry generation from low-resolution aerial photographs for building energy modeling | |
| CN109543616B (en) | Method, device, equipment and medium for evaluating damage of target object after earthquake | |
| CN108304809B (en) | Near real-time earthquake damage assessment method based on post-earthquake aerial image | |
| CN115409069A (en) | Village and town building identification method, classification method, device, electronic equipment and medium | |
| Kaur et al. | Convolutional neural network based hurricane damage detection using satellite images | |
| CN110956412A (en) | Flood dynamic assessment method, device, medium and equipment based on real-scene model | |
| Dogan et al. | Detection of damages caused by earthquake and reinforcement corrosion in RC buildings with Deep Transfer Learning | |
| Torkko et al. | How to best map greenery from a human perspective? Comparing computational measurements with human perception | |
| KR20200094897A (en) | Apparatus and method for providing diagnostic report on building defect | |
| Kanwar et al. | Damage detection for framed RCC buildings using ANN modeling | |
| Vervoort | Application of neural networks to rate of spread estimation in shrublands | |
| JP2020060810A (en) | System for supporting inspections of infrastructure facilities, method and program for supporting inspections of infrastructure facilities | |
| Torres et al. | Using remote sensing for exposure and seismic vulnerability evaluation: is it reliable? | |
| Shi et al. | An innovative approach to determine building window-to-wall ratios for urban energy simulation | |
| Hu et al. | Framework for prioritizing geospatial data processing tasks during extreme weather events | |
| Mishra et al. | Simulation in Architectural Research | |
| Pietro et al. | Automatic identification of residential building features using machine learning techniques | |
| Kang | From rupture to recovery: Integrating probabilistic building performance assessment, decision-making and socioeconomic vulnerability to model post-earthquake housing recovery | |
| CN113505952B (en) | Masonry structure building group safety assessment method based on geometric and performance similarity | |
| US20230027774A1 (en) | Smart real estate evaluation system | |
| Muin | Structural Health Monitoring Framework Using Acceleration Data and Machine Learning Techniques | |
| Nagila et al. | An Innovative Approach of CNN-BiGRU Based Post-Earthquake Damage Detection of Reinforced Concrete for Frame Buildings | |
| CN116151615A (en) | Earthquake disaster personnel risk assessment method, system and electronic equipment |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |