Community extreme disaster resilience generation method and system based on street view images

By calculating feature type proportions and processing of regression model of street scene images, the problem of time-consuming and errors in the judgment of resilience of disaster-stricken communities in the existing technology is solved, and a rapid and accurate disaster recovery assessment is achieved, supporting refined post-disaster assistance is supported. .

CN117973922BActive Publication Date: 2024-11-12SHENZHEN UNIV
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202311871848.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-29
Publication Date
2024-11-12
Estimated Expiration
2043-12-29

AI Technical Summary

Technical Problem

In the prior art, when judging the recovery ability of extreme disaster scenarios for disaster-affected communities, a large amount of data needs to be processed, which is time-consuming and labor-intensive and prone to errors.

Method used

By obtaining the street view image group of the target area, the proportion of the image feature type is calculated, and inputting it into the trained regression model to output resilience information. This method constructs training data based on mobile phone signaling spatiotemporal activity data and training street scene image groups, uses semantic segmentation model to process street scene images, calculates the proportion of feature types of each community, and obtains resilience information through the regression model.

Benefits of technology

It has achieved a rapid and accurate understanding of community extreme disaster scenarios and supported a large-scale refined post-disaster assistance and disaster prevention and mitigation space governance activities.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117973922B_ABST
    Figure CN117973922B_ABST
Patent Text Reader

Abstract

The present invention discloses a method and system for generating community extreme disaster resilience based on street view images, wherein the method comprises: obtaining a street view image group in a target area within a target time, and calculating the proportion of image feature types in the street view image group; inputting the proportion of image feature types in the street view image group into a trained regression model, and outputting resilience information; wherein the regression model is trained by training data, and the training data is constructed based on mobile phone signaling spatiotemporal activity data and a training street view image group. The present invention can quickly and accurately obtain resilience information through the method, thereby facilitating relevant personnel to carry out governance activities based on the resilience information.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to next-generation information technology, and in particular to a method, system, terminal, and storage medium for generating community resilience to extreme disasters based on street view images. Background Technology

[0002] Extreme disasters currently have a very serious impact on society and regions. After an extreme disaster occurs, in the subsequent large-scale and refined disaster relief and disaster prevention and mitigation spatial governance activities, it is extremely important to assess the resilience of the affected communities in extreme disaster scenarios.

[0003] However, current assessments of the resilience of affected communities in extreme disaster scenarios mostly rely on manual processing to obtain resilience information. This process involves handling large amounts of data, and the current methods of obtaining resilience information are time-consuming and labor-intensive, which can easily lead to errors in assessing the resilience of communities in extreme disaster scenarios.

[0004] Therefore, existing technologies still need improvement and development. Summary of the Invention

[0005] The main objective of this invention is to provide a method, system, terminal, and storage medium for generating community resilience in extreme disaster scenarios based on street view images. This aims to solve the problem that in the prior art, when assessing the resilience of affected communities in extreme disaster scenarios, a large amount of data needs to be processed, and the current methods of acquiring resilience information are time-consuming and labor-intensive, which can easily lead to errors in assessing the resilience of communities in extreme disaster scenarios.

[0006] To achieve the aforementioned objective, a first aspect of the present invention provides a method for generating community resilience to extreme disasters based on street view images, wherein the method comprises:

[0007] Acquire a group of street view images in a target area within a target time period, and calculate the proportion of image feature types in the street view image group;

[0008] The proportion of image feature types in the street view image group is input into the trained regression model, and the resilience information is output.

[0009] Specifically, the regression model is trained using training data, which is constructed based on mobile phone signaling spatiotemporal activity data and a training street view image set.

[0010] Optionally, the training steps of the regression model include:

[0011] Acquire training street view image sets and mobile signaling spatiotemporal activity data for a preset number of regions;

[0012] The training data is constructed based on training street view image sets and mobile signaling spatiotemporal activity data for each region;

[0013] The regression model is trained based on the training data to obtain the trained regression model.

[0014] Optionally, the step of generating training data based on street view image sets and mobile signaling spatiotemporal activity data for each region includes:

[0015] A training disaster recovery cycle is generated based on the spatiotemporal activity data of mobile signaling in each of the aforementioned regions;

[0016] Based on the training street view image group for each region, the proportion of image feature types in each training street view image group is calculated to obtain the training proportion data for each region.

[0017] The training data is constructed based on the training disaster recovery cycle and corresponding training ratio data for each region.

[0018] Optionally, the step of generating a training disaster recovery cycle based on mobile phone signaling spatiotemporal activity data for each of the regions includes:

[0019] The mobile phone signaling spatiotemporal activity data for each region is preprocessed to obtain multiple sets of coordinate data for each region, wherein the coordinate data includes the user's location information;

[0020] Based on multiple sets of coordinate data for each region, obtain the daily activity communities of users in each region;

[0021] Obtain the population baseline of the active community in each region within a third preset time period, and calculate the population change rate of each region based on the population baseline and multiple sets of coordinate data. Obtain the training disaster recovery cycle of each region based on the population change rate of each region.

[0022] Optionally, the step of calculating the proportion of image feature types in each training street view image group based on each region to obtain training proportion data for each region includes:

[0023] Preprocessing the training street view image set for each region yields the usable training street view image set for each region.

[0024] The available training street view image set for each region is input into the pre-trained semantic segmentation model, and the semantic segmentation result of the available training street view image set for each region is output.

[0025] The proportion of each image feature type in each street view image group is calculated based on the semantic segmentation results of the available street view image groups for each region, thereby obtaining the training proportion data for each region.

[0026] Optionally, the step of calculating the proportion of each image feature type in each street view image group based on the semantic segmentation results of each available street view image group for each region includes:

[0027] Based on each semantic segmentation result, obtain the number of all image features in each training street scene image group and the number of each image feature type in each direction;

[0028] The proportion of each image feature type in each training street view image group is calculated based on the total number of all image features in each training street view image group and the number of each image feature type in each direction.

[0029] Optionally, the step of acquiring a group of street view images in a target area within a target time period and calculating the proportion of image feature types in the group of street view images includes:

[0030] Acquire a group of street view images of the target area within a target time period;

[0031] The street view image set is preprocessed to obtain a usable street view image set;

[0032] The available street view image set is input into the trained semantic segmentation model, and the current semantic segmentation result is output.

[0033] The proportion of each image feature type in the street view image group is calculated based on the current semantic segmentation result.

[0034] A second aspect of the present invention provides a community extreme disaster resilience generation system based on street view images, wherein the community extreme disaster resilience generation system based on street view images includes:

[0035] The data acquisition module is used to acquire a group of street view images in a target area within a target time period, and to calculate the proportion of image feature types in the group of street view images;

[0036] The result output module is used to input the proportion of image feature types in the street view image group into the trained regression model and output the resilience information.

[0037] A third aspect of the present invention provides a terminal, the terminal including a memory, a processor, and a community extreme disaster resilience generation program based on street view images stored in the memory and executable on the processor, wherein when the community extreme disaster resilience generation program based on street view images is executed by the processor, it implements any of the steps of the community extreme disaster resilience generation method based on street view images.

[0038] A fourth aspect of the present invention provides a computer-readable storage medium storing a community extreme disaster resilience generation program based on street view images, wherein when the community extreme disaster resilience generation program based on street view images is executed by a processor, it implements the steps of any one of the methods for generating community extreme disaster resilience based on street view images.

[0039] As can be seen from the above, in the present invention, a group of street view images in a target area within a target time period is acquired, and the proportion of image feature types in the street view image group is calculated; the proportion of image feature types in the street view image group is input into a trained regression model, and resilience information is output; wherein, the regression model is trained through training data, and the training data is constructed based on mobile phone signaling spatiotemporal activity data and training street view image group.

[0040] Compared to existing technologies, assessing the resilience of communities in extreme disaster scenarios is time-consuming, labor-intensive, and lacks a consistent assessment method, easily leading to errors in the assessment. This invention addresses this problem by using mobile signaling spatiotemporal activity data to obtain population change rates and disaster recovery cycles in the target area, and correspondingly acquiring pre-disaster street view images of the community. Semantic segmentation is performed on these street view images to classify feature types and calculate the proportion of each feature type in each community. A regression model is then used to process the obtained population change rates and the proportion of each feature type in each community. Training data for the regression model is obtained from the mobile signaling spatiotemporal activity data and the training street view image set. The regression model is trained using this training data, enabling it to derive corresponding resilience information from the street view image set in the target area. This resilience information provides a direct understanding of the current community's resilience in extreme disaster scenarios, allowing for large-scale, refined post-disaster relief and disaster prevention and mitigation spatial governance activities based on the current community's resilience in extreme disaster scenarios. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This is a flowchart illustrating a method for generating community resilience to extreme disasters based on street view images, provided in an embodiment of the present invention.

[0043] Figure 2 This is a schematic diagram of the disaster resilience analysis results of a method for generating community extreme disaster resilience based on street view images provided in an embodiment of the present invention;

[0044] Figure 3 This is a schematic diagram of the regression model training process in a method for generating community resilience to extreme disasters based on street view images, provided in an embodiment of the present invention.

[0045] Figure 4 This is a schematic diagram of the constituent modules of a community extreme disaster resilience generation system based on street view images provided in an embodiment of the present invention;

[0046] Figure 5 This is a block diagram illustrating the internal structure of a terminal provided in an embodiment of the present invention. Detailed Implementation

[0047] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, so as to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0048] It should be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0049] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0050] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0051] As used in this specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to classification." Similarly, the phrases "if determined" or "if classified to [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once classified to [the described condition or event]," or "in response to classification to [the described condition or event]."

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0054] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to user-analyzed data, user-stored data, user-displayed data, etc.) and signals involved in this invention are all information, data and signals authorized by the user or fully authorized by all parties; and the collection, use and processing of relevant information, data and signals comply with the laws, regulations and standards of relevant countries and regions.

[0055] Currently, extreme disasters have a severe impact on society and regions. Following such disasters, large-scale, refined post-disaster relief and disaster prevention and mitigation efforts necessitate assessing the resilience of affected communities in extreme disaster scenarios. However, current assessments of community resilience in extreme disaster scenarios largely rely on manual processing to obtain resilience information. This process involves handling large amounts of data, and the current methods of acquiring resilience information are time-consuming and labor-intensive, easily leading to errors in assessing the resilience of communities in extreme disaster scenarios.

[0056] To address at least one of the aforementioned problems, the present invention provides a method, system, terminal, and storage medium for generating community resilience to extreme disasters based on street view images. Specifically, it involves acquiring a group of street view images in a target area within a target time period and calculating the proportion of image feature types in the street view image group; inputting the proportion of image feature types in the street view image group into a trained regression model to output resilience information; wherein the regression model is trained using training data, and the training data is constructed based on mobile phone signaling spatiotemporal activity data and the training street view image group.

[0057] This invention uses mobile signaling spatiotemporal activity data to obtain population change rates and disaster recovery cycles in a target area. It also acquires pre-disaster street view images of the community, performs semantic segmentation on these images, classifies feature types, and calculates the proportion of each feature type in each community. A regression model is then used to process the obtained population change rates and the proportion of each feature type in each community. Training data for the regression model is obtained from the mobile signaling spatiotemporal activity data and the training street view image set. The regression model is trained using this training data, enabling it to derive corresponding resilience information for the street view image set in the target area. This resilience information provides a direct understanding of the community's resilience in extreme disaster scenarios, allowing for large-scale, refined post-disaster relief and disaster prevention and mitigation spatial governance activities based on the community's resilience in extreme disaster scenarios.

[0058] Exemplary methods

[0059] like Figure 1 As shown, this embodiment of the invention provides a method for generating community resilience to extreme disasters based on street view images. Specifically, the method includes the following steps:

[0060] Step S100: Obtain a group of street view images in the target area within the target time period, and calculate the proportion of image feature types in the group of street view images.

[0061] It should be noted that the target time is a pre-set period of time before the occurrence of an extreme disaster, such as 30 days before the disaster. This time is not limited and can be set by the user according to the actual situation. Extreme disasters include earthquakes, floods, and other disasters.

[0062] Further, the step of acquiring a group of street view images in a target area within a target time period and calculating the proportion of image feature types in the group of street view images includes:

[0063] Acquire a group of street view images of the target area within a target time period;

[0064] The street view image set is preprocessed to obtain a usable street view image set;

[0065] The available street view image set is input into the trained semantic segmentation model, and the current semantic segmentation result is output.

[0066] The proportion of each image feature type in the street view image group is calculated based on the current semantic segmentation result.

[0067] Specifically, within the target time period, a group of street view images of the target area is acquired, preprocessed, and then input into a semantic segmentation model for segmentation. This allows the image feature types of each street view image in each group to be classified, resulting in semantic segmentation results. Based on the semantic segmentation results, the proportion of each image feature type in each street view image group is calculated.

[0068] Acquire all street view image groups in the target area within a first preset time period, and preprocess the street view image groups to obtain usable street view image groups, specifically including:

[0069] Obtain all street view image groups in the target area from the network map within a first preset time period, wherein each street view image group includes images from four directions obtained by the same street view extractor at the same location;

[0070] Remove street view image groups that meet the non-target image requirements from all street view image groups, and obtain usable street view image groups based on the remaining street view image groups.

[0071] Specifically, pre-disaster street view images of each community are extracted. The street view image groups are sourced from online maps. During the extraction of street view image groups, four images are extracted relative to the direction of travel of each street view extraction camera, and these four images are combined into one street view image group. When removing street view image groups that meet the requirements of non-target images from all the street view image groups, the selection criteria for non-target images are that the street view images in the street view image group are images from winter months (December-February) and before a preset year. In this application, the preset year is preferably set to 2015. At the same time, street view image groups containing street view images defined as urban highways in OpenStreetMap are excluded, as are street view image groups containing street view images that do not contain sidewalk elements. After removing the above street view image groups, usable street view image groups are obtained.

[0072] Furthermore, the current semantic segmentation result includes street view image feature types, which are one of 13 street view image feature types, including buildings, windows or facade openings, vehicular roads, pedestrian roads, street furniture, trees, grass and shrubs, sky, nature, pedestrians, four-wheeled non-freight motor vehicles, four-wheeled freight motor vehicles, and two-wheeled vehicles.

[0073] Specifically, semantic segmentation is performed on the street scene images in the street scene image group. The semantic segmentation model is obtained using the ResNet18dilated+PPM_deepsup architecture. In this application, the semantic segmentation model segments 13 street scene image feature types, namely buildings, windows or facade openings, vehicular roads, pedestrian roads, street furniture, trees, grass and shrubs, sky, nature, pedestrians, four-wheeled non-freight motor vehicles, four-wheeled freight motor vehicles, and two-wheeled vehicles. See Table 1 below for details.

[0074]

[0075] Table 1. Feature and Semantic Comparison Table

[0076] In this application, training image data for a semantic segmentation model is obtained in advance. The training image data includes the features listed in Table 1 above. Each image in the training image data is labeled to form a training dataset. The semantic segmentation model is trained based on the training dataset. When the number of training iterations reaches a preset number or the semantic segmentation model accuracy reaches a preset accuracy, training is stopped, and the semantic segmentation model obtained from the last training iteration is output to obtain the trained semantic segmentation model. The training image data for the semantic segmentation model is also obtained from street view images in a network map.

[0077] The available street view image set obtained in this application is input into the trained semantic segmentation model, so that the current semantic segmentation result can be output.

[0078] Furthermore, the step of calculating the proportion of each image feature type in the street view image group based on the current semantic segmentation results includes:

[0079] Based on the semantic segmentation results, obtain the number of all image features in the street view image group and the number of each image feature type in each direction;

[0080] Calculate the proportion of each image feature type in the street view image group based on the total number of all image features in the street view image group and the number of each image feature type in each direction.

[0081] Specifically, in this application, the number of all image features in the street view image group is obtained. and the number of each image feature type in each direction Where j represents the direction, and i represents the number of images classified as image feature type i at the j-direction. The proportion of each image feature type in each street view image group is calculated by the following formula, as shown in the following formula (1):

[0082]

[0083] Wherein, SVFi is the proportion of each image feature type in the street view image group; the proportion of each image feature type in all street view image groups can be obtained by calculation.

[0084] Step S200: Input the proportion of image feature types in the street view image group into the trained regression model and output the resilience information.

[0085] It should be noted that the regression model is trained using training data, which is constructed based on mobile signaling spatiotemporal activity data and a training street view image set. In this application, the regression model is the LASSO regression model provided by Pyth on packagescikit-learn, i.e., lasso regression. A penalty function is generated to compress the coefficients of the variables in the regression model, thereby preventing overfitting and solving the problem of severe multicollinearity. In this application, the proportion of image feature types in the obtained street view image set is input into the trained regression model, which outputs the corresponding disaster recovery cycle, i.e., resilience information. Through the resilience information, the resilience of the current community under extreme disaster scenarios can be intuitively understood, thereby enabling large-scale and refined post-disaster relief and disaster prevention and mitigation spatial governance activities based on the resilience of the current community under extreme disaster scenarios.

[0086] Specifically, in one embodiment of this application, the restoring force of multiple target regions is acquired and represented graphically to obtain the following: Figure 2 The diagram shown illustrates the level of resilience in different regions, providing a convenient way to understand the resilience levels in these regions.

[0087] Furthermore, the training steps of the regression model include:

[0088] Acquire training street view image sets and mobile signaling spatiotemporal activity data for a preset number of regions;

[0089] The training data is constructed based on training street view image sets and mobile signaling spatiotemporal activity data for each region;

[0090] The regression model is trained based on the training data to obtain the trained regression model.

[0091] Specifically, a regression model is trained based on the training data, and the model coefficients are updated. Training ends when the number of training iterations reaches a preset requirement or when the model accuracy obtained from the test set reaches the set accuracy requirement, and the trained regression model is output. The preset number refers to the pre-set dimensions of the training data. The training street view image set is also obtained from a network map. Mobile signaling spatiotemporal activity data is the communication data between mobile phone users and transmitting base stations or micro-stations. This data is used to determine the user's spatial location and can relatively accurately record the spatiotemporal trajectory of population movement. Furthermore, the mobile signaling spatiotemporal activity data obtained in this application refers to the time following the occurrence of an extreme disaster in each region. In this application, the preset number of training street view image sets and mobile signaling spatiotemporal activity data for each region refer to regions where extreme disasters have occurred. The training street view image sets are taken at a time prior to the occurrence of an extreme disaster in each region, such as 30 days before the disaster. This time is not limited and can be set by the user according to actual circumstances.

[0092] Furthermore, such as Figure 3 As shown in this application, the population change rate during the disaster in each region is calculated to obtain the training disaster recovery cycle, and a training street view image group is collected and the proportion of image feature types is calculated. Thus, the regression model is trained by training the disaster recovery cycle and the proportion of image feature types.

[0093] Furthermore, the step of generating training data based on street view image sets and mobile signaling spatiotemporal activity data for each region includes:

[0094] A training disaster recovery cycle is generated based on the spatiotemporal activity data of mobile signaling in each of the aforementioned regions;

[0095] Based on the training street view image group for each region, the proportion of image feature types in each training street view image group is calculated to obtain the training proportion data for each region.

[0096] The training data is constructed based on the training disaster recovery cycle and corresponding training ratio data for each region.

[0097] Specifically, when constructing the training data, the same operations as those used in processing the street view image group in the target area are applied to the training street view image group to obtain the training proportion data for each region. Furthermore, a training disaster recovery cycle is generated based on the spatiotemporal activity data of mobile phone signaling for each region. The training proportion data for each region is used as a data sample, and the corresponding training disaster recovery cycle is used as the label for the data sample, thereby constructing the training data.

[0098] Furthermore, the step of generating a training disaster recovery cycle based on mobile phone signaling spatiotemporal activity data for each of the regions includes:

[0099] The mobile phone signaling spatiotemporal activity data for each region is preprocessed to obtain multiple sets of coordinate data for each region, wherein the coordinate data includes the user's location information;

[0100] Based on multiple sets of coordinate data for each region, obtain the daily activity communities of users in each region;

[0101] Obtain the population baseline of the active community in each region within a third preset time period, and calculate the population change rate of each region based on the population baseline and multiple sets of coordinate data. Obtain the training disaster recovery cycle of each region based on the population change rate of each region.

[0102] Specifically, micro-level population distribution data with geographic coordinates within the target area during a disaster is obtained through mobile phone signaling. More specifically, this involves acquiring spatiotemporal activity data of mobile phone signaling within the target area. After acquiring the spatiotemporal activity data, its format is preprocessed in Python to standardize it. This data includes longitude, latitude, timestamp, and user identification code. A user's spatiotemporal activity data is a sequence of location points ordered by time. Device information using unlabeled IDs is removed. To standardize the reporting cycle of various devices and protect residents' location information, this application sets the collection interval for mobile phone signaling spatiotemporal activity data to once every 15 minutes, and stores it. During storage, the daily mobile phone signaling spatiotemporal activity data is stored as a set of coordinate data. Within each set of coordinate data, storage space is divided based on the community unit in the current area to avoid leaking the detailed and precise coordinates of individual residents. In addition, based on the geographic coordinates of highways and main roads, a data cleaning range was set with an extension of 25m on both sides of the road centerline. In Python, mobile phone signaling spatiotemporal activity data within the data cleaning range were deleted through system settings. Multiple sets of coordinate data were obtained through the above preprocessing.

[0103] In this application, when obtaining a user's daily activity communities, argmax is used to represent each mobile phone user's main activity communities for that day. argmax is the variable or set of variables that maximizes the function's value within its range; in this application, it represents the community space most frequently visited by the user each day. (Using vectors...)<p,tp> Each ping represents a coordinate data point within a set of coordinate data. The user's activity area can then be represented by the following formula: Hu, Where Hu,dn represents the primary activity community space of user u on day n, Pu represents the coordinates of all spatial communities visited by user u (i.e., coordinate data), Ru represents all sampled time points, i represents a specific device, and tp represents the spatial coordinates of a device at time tp. For each set of coordinate data, for Monday to Friday, this application uses data from 9 PM to 7 AM as the coordinate data for the day containing 9 PM to ensure that the data reflects the user's primary living community.

[0104] Furthermore, the steps to obtain the population baseline of the active communities in each region within a third preset time period include:

[0105] Obtain the average population of the affected community in the 15 days prior to the disaster;

[0106] The average population size is used as the population baseline.

[0107] Specifically, when calculating the population change rate, a population baseline needs to be set. In this application, the average population of each active community in the 15 days prior to the disaster is calculated as the population baseline for that active community, denoted by Ra.

[0108] The population baseline of the active communities within 15 days prior to the disaster is obtained. Based on the population baseline and multiple sets of coordinate data, the daily population change rate is calculated. Specifically, the population change rate of each active community is calculated daily after the occurrence of the extreme disaster, using the formula: R = (Rdn - Ra) / Ra. Here, Rdn represents the difference between the steady-state and disturbed user numbers on day n, and Ra is the average number of users in each community on the pre-disaster steady-state day. The calculated population change rate indicates the time it takes for the population to recover to the pre-disaster steady-state value from the date of the natural disaster, thus determining the time it takes for the population of each active community to return to normal. Therefore, the corresponding population change rate can characterize the training disaster recovery cycle of each region.

[0109] Further, the step of calculating the proportion of image feature types in each training street view image group based on each region to obtain the training proportion data for each region includes:

[0110] Preprocessing the training street view image set for each region yields the usable training street view image set for each region.

[0111] The available training street view image set for each region is input into the pre-trained semantic segmentation model, and the semantic segmentation result of the available training street view image set for each region is output.

[0112] The proportion of each image feature type in each street view image group is calculated based on the semantic segmentation results of the available street view image groups for each region, thereby obtaining the training proportion data for each region.

[0113] Specifically, street view image groups that meet the non-target image requirements are removed from all the training street view image groups. The non-target image requirements are that the street view images in the training street view image groups are images from winter months (December-February) and earlier than a preset year, where the preset year is preferably set to 2015 in this application. At the same time, street view image groups containing street view images defined as urban highways in OpenStreetMap are excluded, as are street view image groups containing street view images that do not contain sidewalk features. After removing the above training street view image groups, usable training street view image groups are obtained.

[0114] The available training street view image set for each region is input into the trained semantic segmentation model, thereby outputting the semantic segmentation result of the available training street view image set for each region.

[0115] Further, the step of calculating the proportion of each image feature type in each street view image group based on the semantic segmentation results of each available street view image group for each region includes:

[0116] Based on each semantic segmentation result, obtain the number of all image features in each training street scene image group and the number of each image feature type in each direction;

[0117] The proportion of each image feature type in each training street view image group is calculated based on the total number of all image features in each training street view image group and the number of each image feature type in each direction.

[0118] Specifically, the operation of calculating the proportion of each image feature type in each training street view image group is the same as that of calculating the proportion of each image feature type in the street view image group, and therefore will not be repeated here.

[0119] As can be seen from the above, compared with existing technologies, the current methods for assessing the resilience of communities in extreme disaster scenarios are time-consuming, labor-intensive, and lack a consistent assessment method, which can easily lead to errors in the assessment of community resilience in extreme disaster scenarios. This invention uses mobile phone signaling spatiotemporal activity data to obtain the population change rate and disaster recovery cycle in the target area, and correspondingly obtains pre-disaster street view images of the community. The street view images are then semantically segmented, the feature element types are classified, and the proportion of feature types in each community is calculated. A regression model is then used to process the obtained population change rate and the proportion of feature types in each community. Training data for training the regression model is obtained from the mobile phone signaling spatiotemporal activity data and the training street view image set. The regression model is trained using the training data, enabling it to derive corresponding resilience information for the street view image set in the target area. The obtained resilience information provides an intuitive understanding of the current community's resilience in extreme disaster scenarios, allowing for large-scale and refined post-disaster relief and disaster prevention and mitigation spatial governance activities based on the current community's resilience in extreme disaster scenarios.

[0120] Exemplary device

[0121] like Figure 4 As shown in the figure, corresponding to the method for generating community resilience to extreme disasters based on street view images, this embodiment of the invention also provides a system for generating community resilience to extreme disasters based on street view images, the system comprising:

[0122] The data acquisition module 21 is used to acquire a group of street view images in a target area within a target time period, and to calculate the proportion of image feature types in the group of street view images.

[0123] The result output module 22 is used to input the proportion of image feature types in the street view image group into the trained regression model and output the resilience information.

[0124] It should be noted that the specific structure and implementation of the community extreme disaster resilience generation system based on street view images and its various modules or units can be referred to the corresponding description in the method embodiment, and will not be repeated here.

[0125] It should be noted that the division of the modules in the community extreme disaster resilience generation system based on street view images is not unique and is not intended as a specific limitation.

[0126] Based on the above embodiments, the present invention also provides a terminal, the principle block diagram of which can be as follows: Figure 5As shown. The terminal includes a processor 10, a memory 20, a network interface, and a display screen 30 connected via a system bus. In one embodiment, when the processor 10 executes a community extreme disaster resilience generation program 40 based on street view images stored in the memory 20, the following steps are implemented:

[0127] Acquire a group of street view images in a target area within a target time period, and calculate the proportion of image feature types in the street view image group;

[0128] The proportion of image feature types in the street view image group is input into the trained regression model, and the resilience information is output.

[0129] Specifically, the regression model is trained using training data, which is constructed based on mobile phone signaling spatiotemporal activity data and a training street view image set.

[0130] Optionally, the training steps of the regression model include:

[0131] Acquire training street view image sets and mobile signaling spatiotemporal activity data for a preset number of regions;

[0132] The training data is constructed based on training street view image sets and mobile signaling spatiotemporal activity data for each region;

[0133] The regression model is trained based on the training data to obtain the trained regression model.

[0134] Optionally, the step of generating training data based on street view image sets and mobile signaling spatiotemporal activity data for each region includes:

[0135] A training disaster recovery cycle is generated based on the spatiotemporal activity data of mobile signaling in each of the aforementioned regions;

[0136] Based on the training street view image group for each region, the proportion of image feature types in each training street view image group is calculated to obtain the training proportion data for each region.

[0137] The training data is constructed based on the training disaster recovery cycle and corresponding training ratio data for each region.

[0138] Optionally, the step of generating a training disaster recovery cycle based on mobile phone signaling spatiotemporal activity data for each of the regions includes:

[0139] The mobile phone signaling spatiotemporal activity data for each region is preprocessed to obtain multiple sets of coordinate data for each region, wherein the coordinate data includes the user's location information;

[0140] Based on multiple sets of coordinate data for each region, obtain the daily activity communities of users in each region;

[0141] Obtain the population baseline of the active community in each region within a third preset time period, and calculate the population change rate of each region based on the population baseline and multiple sets of coordinate data. Obtain the training disaster recovery cycle of each region based on the population change rate of each region.

[0142] Optionally, the step of calculating the proportion of image feature types in each training street view image group based on each region to obtain training proportion data for each region includes:

[0143] Preprocessing the training street view image set for each region yields the usable training street view image set for each region.

[0144] The available training street view image set for each region is input into the pre-trained semantic segmentation model, and the semantic segmentation result of the available training street view image set for each region is output.

[0145] The proportion of each image feature type in each street view image group is calculated based on the semantic segmentation results of the available street view image groups for each region, thereby obtaining the training proportion data for each region.

[0146] Optionally, the step of calculating the proportion of each image feature type in each street view image group based on the semantic segmentation results of each available street view image group for each region includes:

[0147] Based on each semantic segmentation result, obtain the number of all image features in each training street scene image group and the number of each image feature type in each direction;

[0148] The proportion of each image feature type in each training street view image group is calculated based on the total number of all image features in each training street view image group and the number of each image feature type in each direction.

[0149] Optionally, the step of acquiring a group of street view images in a target area within a target time period and calculating the proportion of image feature types in the group of street view images includes:

[0150] Acquire a group of street view images of the target area within a target time period;

[0151] The street view image set is preprocessed to obtain a usable street view image set;

[0152] The available street view image set is input into the trained semantic segmentation model, and the current semantic segmentation result is output.

[0153] The proportion of each image feature type in the street view image group is calculated based on the current semantic segmentation result.

[0154] The terminal's processor provides computing and control capabilities. The terminal's memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and a community extreme disaster resilience generation program based on street view images. The internal memory provides an environment for the operation of the operating system and the community extreme disaster resilience generation program based on street view images in the non-volatile storage medium. The terminal's network interface is used for communication with external terminals via a network connection. When the community extreme disaster resilience generation program based on street view images is executed by the processor, it implements the steps of any of the community extreme disaster resilience generation methods based on street view images. The terminal's display screen can be a liquid crystal display (LCD) or an e-ink display.

[0155] Those skilled in the art will understand that Figure 5 The schematic diagram shown is merely a partial structural diagram related to the present invention and does not constitute a limitation on the terminal to which the present invention is applied. A specific terminal may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0156] In one embodiment, a terminal is provided, the terminal including a memory, a processor, and a community extreme disaster resilience generation program based on street view images stored in the memory and executable on the processor. When the community extreme disaster resilience generation program based on street view images is executed by the processor, it implements the steps of any of the community extreme disaster resilience generation methods based on street view images provided in the embodiments of the present invention.

[0157] This invention also provides a computer-readable storage medium storing a community extreme disaster resilience generation program based on street view images. When executed by a processor, the community extreme disaster resilience generation program based on street view images implements the steps of any of the community extreme disaster resilience generation methods based on street view images provided in this invention.

[0158] It should be understood that the sequence number of each step in the embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0159] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the division of the functional units and modules is only described as an example. In practical applications, the functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this invention. The specific working process of the units and modules in the device can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0160] In the embodiments described, each embodiment has its own emphasis. For parts not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0161] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0162] In the embodiments provided by this invention, it should be understood that the disclosed system / terminal device and method can be implemented in other ways. For example, the system / terminal device embodiments described above are merely illustrative. For instance, the division of modules or units is merely a logical functional division, and in actual implementation, it can be divided in other ways. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0163] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0164] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not mean that the essence of the corresponding technical solutions deviates from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for generating community extreme disaster resilience based on street view images, characterized in that: The method for generating community extreme disaster resilience based on street view images comprises: Acquire a street view image group in a target area within a target time, and calculate the proportion of image feature types in the street view image group; Inputting the proportion of image feature types in the street view image group into the trained regression model, and outputting resilience information; The regression model is trained by training data, and the training data is constructed based on mobile phone signaling spatiotemporal activity data and a training street view image group; The training steps of the regression model include: Obtain training street view image groups and mobile phone signaling spatiotemporal activity data for a preset number of regions; Constructing the training data based on the training street view image set and mobile phone signaling spatiotemporal activity data of each region; Training the regression model according to the training data to obtain a trained regression model; The step of generating training data based on the street view image group and mobile phone signaling spatiotemporal activity data of each region comprises: generating training disaster recovery cycles based on the spatiotemporal activity data of cell phone signaling in each of said regions; According to the training street view image group of each of the regions, calculating the ratio of image feature types of each of the training street view image groups to obtain training ratio data of each of the regions; constructing the training data according to the training disaster recovery cycle and corresponding training ratio data of each of the regions; The step of generating a training disaster recovery cycle based on the spatiotemporal activity data of mobile phone signaling in each of the regions comprises: Preprocessing the mobile phone signaling spatiotemporal activity data of each of the regions to obtain multiple sets of coordinate data of each region, wherein the coordinate data includes the user's location information; Based on multiple sets of coordinate data for each region, obtain the daily activity communities of users in each region; Obtaining a population baseline of the active community in each region within a third preset time, and calculating a population change rate of each region based on the population baseline of each region and the plurality of sets of coordinate data, and obtaining a training disaster recovery period of each region based on the population change rate of each region; The step of calculating the ratio of image feature types of each training street view image group according to each training street view image group of the region to obtain training ratio data of each region comprises: Preprocessing the training street view image group of each of the regions to obtain an available training street view image group of each of the regions; Inputting the available training street view image group of each of the regions into a pre-trained semantic segmentation model, and outputting the semantic segmentation result of the available training street view image group of each of the regions; Calculate the proportion of each image feature type in each street view image group according to the semantic segmentation result of the available street view image group in each of the regions, and obtain training proportion data for each of the regions; The step of calculating the proportion of each image feature type in each street view image group according to the semantic segmentation result of each street view image group available in the area comprises: Acquire the number of all image features in each of the training street view image groups and the number of each type of image features in each direction according to each of the semantic segmentation results; According to the number of all image features in each of the training street view image groups and the number of each image feature type in each direction, the ratio of each image feature type in each of the training street view image groups is calculated.

2. The method for generating community extreme disaster resilience based on street view images according to claim 1, characterized in that: The step of acquiring a street view image group in a target area within a target time and calculating the proportion of image feature types in the street view image group comprises: Acquire a group of street view images in the target area within a target time; Preprocessing the street view image group to obtain a usable street view image group; Inputting the available street view image group into the trained semantic segmentation model, and outputting the current semantic segmentation result; The proportion of each image feature type in the street view image group is calculated according to the current semantic segmentation result.

3. A community extreme disaster resilience generation system based on street view images, characterized in that: The community extreme disaster resilience generation system based on street view images implements the steps of the community extreme disaster resilience generation method based on street view images as claimed in claim 1, and the community extreme disaster resilience generation system based on street view images includes: A data acquisition module, used to acquire a street view image group in a target area within a target time, and calculate the proportion of image feature types in the street view image group; The result output module is used to input the proportion of image feature types in the street view image group into the trained regression model and output the resilience information.

4. A terminal, characterized in that: The terminal includes a memory, a processor, and a community extreme disaster resilience generation program based on street view images that is stored in the memory and can be run on the processor. When the community extreme disaster resilience generation program based on street view images is executed by the processor, the steps of a community extreme disaster resilience generation method based on street view images as described in any one of claims 1-2 are implemented.

5. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program for generating community extreme disaster resilience based on street view images. When the program for generating community extreme disaster resilience based on street view images is executed by a processor, the steps of a method for generating community extreme disaster resilience based on street view images as described in any one of claims 1-2 are implemented.

Citation Information

Patent Citations

  • Urban traffic state sensing method based on resident travel structure

    CN114358990A

  • Urban disaster toughness fine measurement and calculation method based on space-time activity analysis

    CN116934073A

  • Apparatus for analyzing ground displacement safe neighborhood using land use and space analysis and method thereof

    KR102118335B1

  • Landslide recognition method based on laplacian pyramid remote sensing image fusion

    US11521377B1

  • Performing semantic segmentation of 3D data using deep learning

    US11694333B1