Database management system based on building remote sensing images

Abstract

The invention relates to the technical field of data management, and discloses a database management system based on building remote sensing images. Comprising the following steps: identifying a target image from a building remote sensing image, dividing the target image into an image set, correspondingly importing the image set into a data layer of a library architecture, generating an image database, constructing an associated channel with an on-off point, configuring on-off conditions on the associated channel, analyzing a query demand from request data, and storing the query demand in the database. Selecting a corresponding query mode; compared with the prior art, the method has the advantages that the image association and query channels can be constructed among different data levels in the image database, and meanwhile, the on-off points configured with the on-off conditions are combined, so that the target images in the image database can be dynamically associated through the association channels on the basis of keeping independent, and the dynamic association efficiency is improved. Therefore, the dynamic storage association effect of the target image in the image database is realized, and the diversified query requirements of the image database are met.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and more specifically, to a database management system based on building remote sensing images. Background Art

[0002] After buildings in a region are damaged by an earthquake disaster, it is necessary to analyze the post-earthquake condition of the building structures in the region. Therefore, remote sensing images of building damage after real historical earthquakes are first collected, and then the building remote sensing images are aggregated into a database to achieve accurate management of building remote sensing images and facilitate users to query and use the required building remote sensing images. Therefore, it is necessary to establish a database to manage building remote sensing images.

[0003] The patent application with reference publication number CN112559534A discloses a remote sensing image data archiving management system and method, including a data integration unit, configured to integrate received aerospace and aviation remote sensing image data, a data archiving unit, configured to classify data, and different types of data are stored in different file forms, a data processing unit, including multiple task processing nodes, each task processing node processes corresponding tasks in parallel, and the tasks include data query, data cleaning, data scheduling, data feature extraction, stereo image global optimization matching, suspicious matching detection, regional network adjustment, digital surface model extraction and splicing, and a data storage unit, configured to store aerospace and aviation remote sensing image raw data and processing results; The existing database management system classifies all building remote sensing images according to specific types and stores the classified building remote sensing images in a unified manner in the database, thereby achieving classified query and management effects of building remote sensing images. For example, in the above-mentioned patent application, it normalizes, uniformly stores, and centrally manages massive multi-source remote sensing image data to achieve overall automated operation effects of data archiving, statistics and other businesses. However, the overall unified storage and management method causes a large number of building remote sensing images with different degrees of association to be mixed together, and it is impossible to ensure that images with different degrees of association always maintain a dynamic independent state. In addition, there is a lack of query mode switching measures between different types of building remote sensing images, which makes the query management process of building remote sensing images not accurate enough, thereby increasing the probability of query failure and query error of building remote sensing images in the database.

[0004] In view of this, the present invention proposes a database management system based on building remote sensing images to solve the above problems. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: a database management system based on building remote sensing images, comprising: An image recognition module is used to receive the building remote sensing image of the target area, extract the basic parameters of the building remote sensing image, and recognize the target image from the building remote sensing image; The image classification module is used to extract the comprehensive image features of the target image, which include image time, geographic coordinates, outline color and building structure. The target image is divided into image sets based on the comprehensive image features. A database construction module is used to construct a library architecture with a data layer, and import the image collection into the data layer of the library architecture to generate an image database; The channel building module is used to build an associated channel with on-off points in the image database, and configure the on-off conditions for controlling the on-off state switching of the on-off points on the associated channel based on the comprehensive image features; The query selection module is used to parse the query requirements from the request data, combine the query requirements with the on-off conditions, control the on-off state of the on-off point, and select the query mode corresponding to the on-off state.

[0006] Furthermore, the basic parameters include the abnormal area ratio and the area integrity; The method for extracting the abnormal area ratio is: Using the preset first length as the longitudinal segmentation standard and the preset second length as the horizontal segmentation standard, the A building remote sensing images are segmented into B sub-areas one by one; The number of all pixels in B sub-regions of A building remote sensing images is counted respectively to obtain B regional values, and the sub-regions whose regional values ​​are less than the regional calibration values ​​are recorded as abnormal regions; The total value of abnormal areas in A building remote sensing images is counted one by one, recorded as abnormal value, and the abnormal value is compared with the total number of sub-areas to obtain the ratio of A abnormal areas; The expression of abnormal area ratio is: ; In the formula, For the The abnormal area ratio of the building remote sensing image, =1,2,...,A, For the The abnormal value of the building remote sensing image, For the The total number of sub-areas in the building remote sensing image.

[0007] Furthermore, the method for extracting regional integrity is: Mark the pixel values ​​of all pixels in the B sub-regions of the A building remote sensing images one by one, and mark the pixel values ​​less than the lower limit of the pixel and the pixel values ​​greater than the upper limit of the pixel as abnormal points; The four inflection points of the B sub-areas are cross-connected to form a diagonal line, and the line segment from the four inflection points on the two diagonals to the intersection of the two diagonals is recorded as a median line, thereby obtaining four median lines; Draw four adjacent pixel circles with equal radius and two circumscribed circles with the midpoints of the four median lines as the centers, and count the number of abnormal points and the total number of pixels in the four pixel circles; The pixel circle whose number of abnormal points is less than four-fifths of the total number of pixel points is recorded as a normal circle, and the number of all normal circles in A building remote sensing images is counted to obtain A normal value; The radius of the pixel circle, the length and width of the building remote sensing image are measured by the scale, and the area of ​​the normal circle is calculated based on the circle area formula. The areas of all normal circles are added together and compared with the total area of ​​the building remote sensing image to obtain A regional completeness. The expression of regional integrity is: ; In the formula, For the The regional completeness of the building remote sensing image, For the The normal value of a building remote sensing image, is the radius of the pixel circle, For the The length of a building remote sensing image, For the The width of a building remote sensing image.

[0008] Furthermore, the target image recognition method is: When the abnormal area ratio of the building remote sensing image is less than the abnormal area safety value, and the area integrity is greater than the area integrity safety value, the building remote sensing image is recorded as the target image, and D target images are obtained; When the abnormal area ratio is greater than or equal to the abnormal area safety value, or the area integrity is less than or equal to the area integrity safety value, the building remote sensing image will not be recorded as the target image.

[0009] Furthermore, the contour color extraction method is: The buildings in the D target images are identified one by one by computer technology, the edge lines of the buildings are identified by edge detection algorithms, and the area inside the edge lines is recorded as the building area; Mark the pixel values ​​of all the pixels in the building area on the D target images respectively, and aggregate the pixels with the same color level to obtain E pixel sets; The number of pixels in the E pixel sets is counted one by one, and the color level with the maximum number of pixels is recorded as D valid levels; The natural language processing technology is used to identify D effective levels of text semantics and digital semantics, and after combining the text semantics and digital semantics head to tail, D contour colors are generated.

[0010] Furthermore, the image set includes a semi-hidden set and an open set, and the division method of the semi-hidden set and the open set is: According to the order of image time, the D target images are numbered in ascending order, with 1 being the first number; After the target images with consistent geographic coordinates are aggregated in ascending order of numbering, a first image group is generated; Among the target images with inconsistent geographic coordinates, target images with consistent outline colors and building structures are aggregated to generate a second image group; After all the target images in the first image group and the second image group are combined, an open set is generated, and after all the remaining target images are summarized, a semi-hidden set is generated.

[0011] Furthermore, the method for generating the image database is as follows: A blank data group with a closed structure is established, and two internal and external combined data layers are constructed in the blank data group, and the internal data layer is recorded as the internal layer, and the external data layer is recorded as the external layer; Count the number of target images in the semi-hidden set and the open set respectively to obtain the hidden value and the open value; Mark the external image bits in the external layer in the same number as the open value, and mark the internal image bits in the internal layer in the same number as the hidden value, to construct the library architecture; Import all target images in the semi-hidden set into the inner image position of the inner layer one by one, forcing the inner layer to be converted into a hidden layer; All target images in the open set are imported into the external image position of the external layer one by one, forcing the external layer to be converted into an open layer, and the library architecture with hidden layers and open layers is recorded as an image database.

[0012] Furthermore, the method of building the associated channel is: A point is randomly marked on the hidden layer and the open layer of the image database, respectively, and recorded as the inner connection point and the outer connection point; Measure the distance between the inner connection point and the outer connection point to obtain the inner and outer connection value; The position of the external connection point on the open layer is continuously adjusted until the internal and external connection values ​​reach the minimum value, and the position of the external connection point is stopped, and one quarter of the minimum value of the internal and external connection value is recorded as the interval value; An associated channel with a bidirectional transmission link is built between the inner connection point and the adjusted outer connection point, and a point corresponding to a length of an interval value from the outer connection point is marked on the associated channel and recorded as a connection point.

[0013] Furthermore, the on / off state includes an on state and an off state: The on-off condition is: when the open layer is not sufficient to meet the query requirements, the control on-off point is switched from the closed state to the open state.

[0014] Furthermore, the on-off state identification method is: Identify the request semantics in the request data through natural speech processing technology, and extract the keywords in the request semantics one by one; Time, coordinates, color and structure are recorded as demand words, keywords in the request data containing any demand word are recorded as valid words, and all valid words are combined to generate query requirements; Count the number of valid words in the query requirements and record it as the effective value; When the effective value is 1 and the effective word is coordinate, the control on-off point is in the closed state; When the effective value is 1 and the effective word is not a coordinate, the control on-off point is in the on state; When the effective value is 2, 3 or 4, the control on / off point is in the open state; The query mode includes independent query mode and combined query mode. The selection method of independent query mode and combined query mode is as follows: When the on / off state of the on / off point is on, select the combined query mode; When the on / off state of the on / off point is off, select the independent query mode.

[0015] The technical effects and advantages of the database management system based on building remote sensing images of the present invention are as follows: (1): By receiving the building remote sensing images of the target area, extracting the basic parameters of the building remote sensing images, and identifying the target images from the building remote sensing images, the quality of the building remote sensing images can be evaluated and analyzed from the two data dimensions of resolution and completeness, thereby screening out the target images that meet the subsequent image database construction, avoiding the negative impact of low-quality building remote sensing images on database construction, and improving the accuracy of image database construction; (2): By constructing a library architecture with a data layer and importing the image collection into the data layer of the library architecture, an image database is generated. Based on the image collection and with the target image as the object, an image database with different data layers can be constructed, so that the image database can store and manage the target image, while also ensuring that the target images with different associations can be orderly aggregated and remain relatively independent. (3): By building an association channel with on-off points in the image database, and taking the comprehensive image features as the configuration basis, configuring the on-off conditions that control the on-off state switching of the on-off points on the association channel, it is possible to build an image association and query channel between different data levels in the image database. At the same time, combined with the on-off points configured with on-off conditions, the opening and closing of the association channel can be accurately controlled, ensuring that the target images in the image database can be dynamically associated through the association channel while maintaining independence, thereby realizing the dynamic storage association effect of the target images in the image database, greatly reducing the probability of query failure and query error, and improving the query accuracy; (4): By parsing the query requirements from the request data, combining the query requirements with the on-off conditions, controlling the on-off status of the on-off points, and selecting the query mode corresponding to the on-off status, the on-off status of the associated channels can be accurately controlled according to the actual query requirements, and the corresponding query mode of the building remote sensing image can be provided to the user end, ensuring that the user end can quickly, accurately and comprehensively query the required building remote sensing images from the image database. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of a database management system based on building remote sensing images provided in the first embodiment of the present invention; Figure 2 A flowchart of a database management method based on building remote sensing images is provided in Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Example 1: Please refer to Figure 1 As shown, the database management system based on building remote sensing images described in this embodiment includes: An image recognition module receives the building remote sensing image of the target area, extracts the basic parameters of the building remote sensing image, and recognizes the target image from the building remote sensing image; The target area refers to the specific geographical location indicated by the building remote sensing image used to construct the database, so that the received building remote sensing image can provide a comprehensive and complete image representation of the geographical location from a bird's-eye view. Since the building remote sensing image is taken from a high altitude bird's-eye view, in this embodiment, the building remote sensing image can be obtained by means of satellite remote sensing photography, drone remote sensing photography, and aerial remote sensing photography, etc., as long as it can meet the photography requirements of the target area; After acquiring the building remote sensing images, it is necessary to conduct preliminary recognition and analysis on the building remote sensing images and extract basic parameters from the building remote sensing images. At this time, the basic parameters can numerically represent the basic image quality of the building remote sensing images and serve as the basis for preliminary recognition and screening of the building remote sensing images. Basic parameters include abnormal area ratio and regional integrity; The abnormal area ratio refers to the ratio between the number of abnormal areas identified as low resolution and the total number of areas in the building remote sensing image, which can be used to indicate the overall resolution of the building remote sensing image. The larger the abnormal area ratio, the greater the ratio between the number of abnormal areas identified as low resolution and the total number of areas in the building remote sensing image, and the worse the quality of the building remote sensing image. The method for extracting the abnormal area ratio is: The preset first length is used as the vertical segmentation standard, and the preset second length is used as the horizontal segmentation standard to segment the A building remote sensing images into B sub-areas one by one; the preset first length and the preset second length are length standards used to segment the building remote sensing images in the vertical direction and the horizontal direction respectively, so as to ensure that the building remote sensing images can be segmented into sub-areas of equal size, so as to ensure that the corresponding areas of all sub-areas are consistent, and to avoid the interference of the inconsistent area sizes of the sub-areas on the subsequent analysis results; The number of all pixels in the B sub-regions of the A building remote sensing image is counted respectively to obtain the B regional values, and the sub-regions whose regional values ​​are less than the regional calibration value are recorded as abnormal regions; the regional calibration value refers to the minimum value of the number of pixels in the sub-region identified as an abnormal region, which can be used as a numerical basis for judging whether a sub-region is an abnormal region; The total value of abnormal areas in A building remote sensing images is counted one by one, recorded as abnormal value, and the abnormal value is compared with the total number of sub-areas to obtain the ratio of A abnormal areas; The expression of abnormal area ratio is: ; In the formula, For the The abnormal area ratio of the building remote sensing image, =1,2,...,A, For the The abnormal value of the building remote sensing image, For the The total number of sub-areas in the building remote sensing image.

[0019] Regional integrity refers to the ratio of the area identified as normal grayscale area to the total area of ​​the region in the building remote sensing image, which can be used to indicate the overall grayscale condition of the building remote sensing image. The greater the regional integrity, the greater the ratio of the area identified as normal grayscale area to the total area of ​​the region, and the better the quality of the building remote sensing image. The method for extracting regional integrity is: The pixel values ​​of all pixels in the B sub-regions of the A building remote sensing images are marked one by one, and the pixel values ​​of which are less than the lower pixel limit value and the pixel values ​​of which are greater than the upper pixel limit value are recorded as abnormal points; the lower pixel limit value and the upper pixel limit value are respectively used to limit the minimum and maximum pixel values ​​of all pixels in the sub-region marked as abnormal points, so as to ensure that the pixel values ​​of which are between the lower pixel limit value and the upper pixel limit value are not abnormal points; specifically, the lower pixel limit value and the upper pixel limit value can represent the brightness of the pixel points corresponding to the sub-region, and the brightness of the pixel points corresponding to the lower pixel limit value is lower, and the brightness of the pixel points corresponding to the upper pixel limit value is higher; The four inflection points of the B sub-areas are cross-connected to form a diagonal line, and the line segment from the four inflection points on the two diagonals to the intersection of the two diagonals is recorded as a median line, thereby obtaining four median lines; With the midpoints of the four median lines as the centers, four adjacent pixel circles with equal radius and two circumscribed circles are drawn, and the number of abnormal points and the total number of pixels in the four pixel circles are counted respectively; by drawing pixel circles, circular areas of equal size can be constructed in the sub-area, and the regular coverage of the circular areas on the pixels can be used to accurately calculate the regional integrity; The pixel circle whose number of abnormal points is less than four-fifths of the total number of pixel points is recorded as a normal circle, and the number of all normal circles in A building remote sensing images is counted to obtain A normal value; The radius of the pixel circle, the length and width of the building remote sensing image are measured by the scale, and the area of ​​the normal circle is calculated based on the circle area formula. The areas of all normal circles are added together and compared with the total area of ​​the building remote sensing image to obtain A regional completeness. The expression of regional integrity is: ; In the formula, For the The regional completeness of the building remote sensing image, For the The normal value of a building remote sensing image, is the radius of the pixel circle, For the The length of a building remote sensing image, For the The width of a building remote sensing image.

[0020] After the basic parameters of the building remote sensing images are analyzed, the building remote sensing images can be preliminarily screened according to the size of the basic parameters, and then the building remote sensing images with high resolution and high image integrity can be identified from the building remote sensing images and recorded as target images; The target image recognition method is: First, the abnormal area ratios of A building remote sensing images are compared with the abnormal area safety values ​​in turn; the abnormal area safety value refers to the maximum value of the abnormal area ratio of the building remote sensing image when it is identified as the target image, which can provide a numerical basis for determining the size of the abnormal area ratio; When the abnormal area ratio of the building remote sensing image is less than the abnormal area safety value, it means that the resolution of the building remote sensing image is high; Then, the regional integrity of the building remote sensing image is compared with the regional integrity safety value; the regional integrity safety value refers to the minimum value of the regional integrity of the building remote sensing image when it is identified as a target image, which can provide a numerical basis for determining the size of the regional integrity; When the regional integrity of the building remote sensing image is greater than the regional integrity safety value, it means that the integrity of the building remote sensing image is high, then the building remote sensing image is recorded as the target image, and D target images are obtained.

[0021] It should be noted that when the abnormal area ratio is greater than or equal to the abnormal area safety value, or the area integrity is less than or equal to the area integrity safety value, it means that the building remote sensing image has low resolution or poor integrity, and the building remote sensing image cannot be used as the target image. Therefore, the screened target image needs to maintain a high level of resolution and integrity in order to provide a high-quality image foundation for the construction of the subsequent database.

[0022] The image classification module extracts the comprehensive image features of the target image and divides the target image into image sets based on the comprehensive image features. Comprehensive image features refer to the specific representation of detailed information such as the type, attributes and structure of the target image, and serve as the basis for classification and management of the target image. Each target image has and only has one set of correct comprehensive image features, so that the comprehensive image features can comprehensively and concisely represent the relevant information in the target image. Comprehensive image features include image time, geographic coordinates, outline color, and building structure; Image time is used to represent the timeline corresponding to the shooting of the target image, which can accurately refer to the shooting time sequence of each target image, thereby providing a reference basis on the timeline for distinguishing the target images; the image time is obtained by querying the shooting time of D target images one by one through the timestamp.

[0023] Geographic coordinates are used to represent the geographic longitude and latitude coordinates corresponding to the target image when it was taken, which can accurately refer to the longitude and latitude position of each target image, thereby providing a reference for the geographical location of the target image; geographic coordinates are obtained by querying the longitude and latitude of D target images one by one through the geographic information management system.

[0024] The outline color is used to represent the corresponding facade color when the target image is shot, which can accurately refer to the facade color of each target image, thereby providing a reference basis for distinguishing the target images on the facade color; The method for extracting contour color is: The buildings in the D target images are identified one by one by computer technology, the edge lines of the buildings are identified by edge detection algorithms, and the area inside the edge lines is recorded as the building area; Mark the pixel values ​​of all the pixels in the building area on the D target images respectively, and summarize the pixel values ​​in the same color level to obtain E pixel sets; the color level is the numerical range between the minimum and maximum values ​​of the pixel values ​​corresponding to different types of colors, which can be used as a basis for judging whether the colors referred to by the pixels are the same color; The number of pixels in the E pixel sets is counted one by one, and the color level with the maximum number of pixels is recorded as D valid levels; The text semantics and digital semantics of D valid levels are identified through natural language processing technology, and the text semantics and digital semantics are combined head to tail to generate D contour colors. The text semantics and digital semantics are used to represent the specific color information and color number in the valid level respectively, and the text semantics and digital semantics are combined to form a specific contour color.

[0025] The building structure is used to represent the corresponding structure type when the target image is taken, that is, to accurately refer to the structure type of each target image, thereby providing a reference for the structure type for distinguishing the target images; the building structure is obtained by querying the building design table after combining the image time, geographic coordinates and outline color of the target image.

[0026] After the comprehensive image features of the target image are extracted, the comprehensive image features can be used as the division standard to orderly and accurately divide the target image into multiple independent image sets, ensuring that each image set contains and only contains target images of the same or similar type; Image sets include semi-hidden sets and open sets; semi-hidden sets refer to target images in an image set that are in a semi-open and semi-hidden state, so that the target images in the image set cannot be directly queried and used; open sets refer to target images in an image set that are in an open state, so that the target images in the image set can be directly queried and used; The division method of semi-hidden set and open set is: According to the order of image time, the D target images are numbered in ascending order, with 1 being the first number; The geographic coordinates of the D target images are compared in order from small to large numbers, and the target images with consistent geographic coordinates are aggregated to generate a first image group; Among the target images with inconsistent geographic coordinates, the outline colors and building structures are compared for consistency, and the target images with consistent outline colors and building structures are aggregated to generate a second image group; After all the target images in the first image group and the second image group are combined, an open set is generated, and after all the remaining target images are summarized, a semi-hidden set is generated.

[0027] It should be noted that the target images in the open set have certain features of the same coordinates, colors or structures, so that the target images in the open set can be associated with each other, so that the target images in the open set can be used directly, while the target images in the semi-hidden set do not have the same and associated features, so that the target images in the semi-hidden set cannot be used directly.

[0028] A database construction module constructs a library architecture with a data layer, and imports the image collection into the data layer of the library architecture to generate an image database; After the target images are classified and divided, a library architecture can be constructed based on the divided target images, so that the library architecture can provide a basic architecture for the database composed of the divided target images, so as to ensure that the subsequently constructed database can effectively and accurately manage all target images, which is convenient for subsequent query and retrieval; In the process of building the library architecture, it is necessary to build a data layer that matches the image collection within the library architecture so that the data layer can provide location constraints for data import for the image collection to ensure that the subsequent image database can match all target images. After the library architecture is constructed, the image collection and the library architecture can be effectively combined and imported, so that the library architecture and the image collection can be combined to construct an image database. At this time, the image database serves as an overall data management library for qualified building remote sensing images in the target area, and can provide a platform for subsequent query and retrieval of related image data. The image database is generated as follows: A blank data group with a closed structure is established, and two internal and external combined data layers are constructed in the blank data group, and the internal data layer is recorded as the internal layer, and the external data layer is recorded as the external layer; the data layer is used to distinguish the locations of building remote sensing images with different degrees of association in the image database, so as to represent the building remote sensing images with different degrees of association with different degrees of openness; Count the number of target images in the semi-hidden set and the open set respectively to obtain the hidden value and the open value; Mark the external image positions with the same number of open values ​​in the external layer, and mark the internal image positions with the same number of hidden values ​​in the internal layer to build a library architecture; the external image positions are used to limit the import positions of building remote sensing images in the open collection; Import all target images in the semi-hidden set into the inner image position of the inner layer one by one, forcing the inner layer to be converted into a hidden layer; the inner image position is used to limit the import position of the building remote sensing image in the hidden set; All target images in the open set are imported into the external image position of the external layer one by one, forcing the external layer to be converted into an open layer, and the library architecture with hidden layers and open layers is recorded as an image database.

[0029] It should be noted that the constructed image database is a data set of building remote sensing images based on all target images in the target area, which can be a set for subsequent user terminals to query and retrieve the required building remote sensing images, and can effectively and accurately manage all target images in the target area as a whole.

[0030] The channel building module builds an associated channel with on-off points in the image database, and configures the on-off conditions for controlling the on-off state switching of the on-off points on the associated channel based on the comprehensive image features; After the image database is constructed, it is necessary to construct association channels between different data layers of the image database for the association transmission and retrieval of target images in different data layers, so as to effectively connect the target images in different data layers of the image database; When building an associated channel, it is necessary to set an on / off point on the associated channel to control the associated channel state, so that the on / off point can be used as a trigger point to control the opening and closing of the associated channel, thereby forming a control point with a protective effect in the image database; The method of building the associated channel is: A point is randomly marked on the hidden layer and the open layer of the image database, respectively, and recorded as the inner connection point and the outer connection point; Measure the distance between the inner connection point and the outer connection point to obtain the inner and outer connection value; The position of the external connection point on the open layer is continuously adjusted until the internal and external connection values ​​reach the minimum value, and the position of the external connection point is stopped, and one quarter of the minimum value of the internal and external connection value is recorded as the interval value; An associated channel with a bidirectional transmission link is built between the inner connection point and the adjusted outer connection point, and a point corresponding to a length of an interval value from the outer connection point is marked on the associated channel and recorded as a connection point.

[0031] It should be noted that when the distance from the external connection point to the internal connection point reaches the minimum value, the distance between the external connection point and the internal connection point is the shortest, and the convenience of associating the target image in the open layer with the hidden layer is the highest. At this time, it is most convenient to perform associative query and retrieval operations on the target image in the open layer and the hidden layer.

[0032] After constructing the associated channel with the on-off point, it is necessary to control the on-off state of the on-off point. At this time, it is necessary to configure the on-off conditions that can control the switching of the on-off point based on the image characteristics of the target image; The on-off state is used to specifically indicate whether the working state of the on-off point is on, and is used as a query restriction condition for the required building remote sensing image in the image database. Specifically, the on-off state includes the on state and the off state; the on state means that the on-off point is in the on state, at which time the target image in the hidden layer and the open layer can be bidirectionally transmitted and queried through the associated channel; the off state means that the on-off point is in the blocked state, at which time the target image in the hidden layer and the open layer cannot be bidirectionally transmitted and queried through the associated channel; The on-off condition is: when the open layer is not sufficient to meet the query requirements, the control on-off point is switched from the closed state to the open state.

[0033] It should be noted that the on / off points of the associated channels are in a closed state when no query management is performed, which ensures that all remote sensing images in the image database remain relatively independent. The open layer is not sufficient to meet the query requirements, which means that the target images in the open layer cannot meet the user's query and management requirements for building remote sensing images. At this time, it is necessary to extend the query range of the required building remote sensing images from the open layer to the hidden layer.

[0034] The query selection module parses the query requirements from the request data, combines the query requirements with the on-off conditions, controls the on-off state of the on-off points, and selects the query mode corresponding to the on-off state; Request data refers to all data that can be used to query, manage and retrieve the required building remote sensing images on the user side, and is used as the original data input to the image database. Since the request data contains many types of requirements and the content is complex, in order to extract concise and accurate requirement data from the request data, it is necessary to parse the query requirements from the request data, so that the query requirements can not only specifically represent the building remote sensing image data and types required in the request data, but also serve as the basis for subsequent determination of satisfaction with the on-off conditions, and after matching and identifying the query requirements with the on-off conditions, the on-off status of the on-off point can be identified; The on / off status identification method is: The request semantics in the request data are identified through natural speech processing technology, and the keywords in the request semantics are extracted one by one; the request semantics is used to concisely express the true meaning of the request data, and the keywords are used to express the specific features represented by the request semantics in words; Record time, coordinates, color, and structure as demand words; Compare the keywords in the request data with the demand words one by one, record the keywords containing any demand word as valid words, and combine all the valid words to generate the query requirements; Count the number of valid words in the query requirements and record it as the effective value; When the effective value is 1 and the effective word is coordinate, the query requirement only needs to query the target image in the open layer of the image database, and there is no need to perform an associated query on the target image in the hidden layer, so the control on-off point is closed; When the effective value is 1 and the effective word is not a coordinate, the query requirement not only needs to query the target image in the open layer of the image database, but also needs to perform an associated query on the target image in the hidden layer, and the control on-off point is in the open state; When the effective value is 2, 3 or 4, the query requirement not only needs to query the target image in the open layer of the image database, but also needs to perform an associated query on the target image in the hidden layer, and the control on-off point is in the open state.

[0035] After the control on-off point is switched to the corresponding on-off state, the query mode corresponding to the on-off state can be selected according to the actual switched on-off state, so that the requested data of the user end can be quickly and accurately queried from the image database to obtain the required building remote sensing image under the restriction of the query mode, thereby realizing the query management operation of the building remote sensing image after the earthquake disaster in the target area, and facilitating the provision of reasonable and accurate data support for the buildings after the earthquake disaster; The query mode includes independent query mode and combined query mode; the independent query mode refers to the query operation of the required building remote sensing image only through the open layer, and the combined query mode refers to the query operation of the required building remote sensing image through the open layer and the hidden layer; The selection method of independent query mode and combined query mode is as follows: When the on / off state of the on / off point is on, it is necessary to perform a combined query operation on the target images in the open layer and the hidden layer through the associated channel, and the combined query mode is selected; When the on / off state of the on / off point is closed, there is no need to perform a combined query operation on the target images in the open layer and the hidden layer through the associated channel, so the independent query mode is selected.

[0036] In this embodiment, by receiving the building remote sensing image of the target area, extracting the basic parameters of the building remote sensing image, and identifying the target image from the building remote sensing image, the quality of the building remote sensing image can be evaluated and analyzed from the two data dimensions of resolution and completeness, so as to screen out the target image that meets the requirements for subsequent image database construction, avoid the negative impact of low-quality building remote sensing images on database construction, and improve the accuracy of image database construction; By extracting the comprehensive image features of the target image and dividing the target image into image sets based on the comprehensive image features, the target images with the same or similar comprehensive image features can be orderly combined to form a relatively independent image set, and provide a data hierarchy basis for the subsequent construction of the image database; By constructing a library architecture with a data layer and importing the image collection into the data layer of the library architecture, an image database is generated. Based on the image collection and with the target image as the object, an image database with different data layers can be constructed, so that the image database can store and manage the target image, and at the same time ensure that the target images with different associations can be orderly aggregated and kept relatively independent. By building an association channel with on-off points in the image database, and taking comprehensive image features as the configuration basis, configuring the on-off conditions for controlling the on-off state switching of the on-off points on the association channel, it is possible to build an image association and query channel between different data levels in the image database. At the same time, combined with the on-off points configured with on-off conditions, it is possible to accurately control the opening and closing of the association channel, ensuring that the target images in the image database can be dynamically associated through the association channel while maintaining independence, thereby realizing the dynamic storage association effect of the target images in the image database, greatly reducing the probability of query failure and query error, and improving the query accuracy; By parsing the query requirements from the request data, combining the query requirements with the on-off conditions, controlling the on-off status of the on-off points, and selecting the query mode corresponding to the on-off status, the on-off status of the associated channels can be accurately controlled according to the actual query requirements, and the corresponding query mode of the building remote sensing image can be provided to the user end, ensuring that the user end can quickly, accurately and comprehensively query the required building remote sensing images from the image database.

[0037] Example 2: Please refer to Figure 2 As shown, the part not described in detail in this embodiment is described in the first embodiment, and a database management method based on building remote sensing images is provided, which is implemented based on a database management system based on building remote sensing images, including: S1: Receive the building remote sensing image of the target area, extract the basic parameters of the building remote sensing image, and identify the target image from the building remote sensing image; S2: Extract the comprehensive image features of the target image, which include image time, geographic coordinates, outline color and building structure. Use the comprehensive image features as the division standard to divide the target image into image sets; S3: construct a library architecture with a data layer, and import the image collection into the data layer of the library architecture to generate an image database; S4: Building an associated channel with on-off points in the image database, and configuring on-off conditions for controlling the on-off state switching of the on-off points on the associated channel based on the comprehensive image features; S5: parse the query requirements from the request data, combine the query requirements with the on-off conditions, control the on-off state of the on-off point, and select a query mode corresponding to the on-off state.

[0038] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A database management system based on building remote sensing images, characterized in that: include: An image recognition module is used to receive the building remote sensing image of the target area, extract the basic parameters of the building remote sensing image, and recognize the target image from the building remote sensing image; The image classification module is used to extract the comprehensive image features of the target image, which include image time, geographic coordinates, outline color and building structure. The target image is divided into image sets based on the comprehensive image features. A database construction module is used to construct a library architecture with a data layer, and import the image collection into the data layer of the library architecture to generate an image database; The channel building module is used to build an associated channel with on-off points in the image database, and configure the on-off conditions for controlling the on-off state switching of the on-off points on the associated channel based on the comprehensive image features; The query selection module is used to parse the query requirements from the request data, combine the query requirements with the on-off conditions, control the on-off state of the on-off point, and select the query mode corresponding to the on-off state.

2. A database management system based on building remote sensing images according to claim 1, characterized in that: Basic parameters include abnormal area ratio and regional integrity; The method for extracting the abnormal area ratio is: Using the preset first length as the longitudinal segmentation standard and the preset second length as the horizontal segmentation standard, the A building remote sensing images are segmented into B sub-areas one by one; The number of all pixels in B sub-regions of A building remote sensing images is counted respectively to obtain B regional values, and the sub-regions whose regional values ​​are less than the regional calibration values ​​are recorded as abnormal regions; The total value of abnormal areas in A building remote sensing images is counted one by one, recorded as abnormal value, and the abnormal value is compared with the total number of sub-areas to obtain the ratio of A abnormal areas; The expression of abnormal area ratio is: ； In the formula, For the The abnormal area ratio of the building remote sensing image, =1,2,...,A, For the The abnormal value of the building remote sensing image, For the The total number of sub-areas in the building remote sensing image.

3. A database management system based on building remote sensing images according to claim 2, characterized in that: The method for extracting regional integrity is: Mark the pixel values ​​of all pixels in the B sub-regions of the A building remote sensing images one by one, and mark the pixel values ​​less than the lower limit of the pixel and the pixel values ​​greater than the upper limit of the pixel as abnormal points; The four inflection points of the B sub-areas are cross-connected to form a diagonal line, and the line segment from the four inflection points on the two diagonals to the intersection of the two diagonals is recorded as a median line, thereby obtaining four median lines; Draw four adjacent pixel circles with equal radius and two circumscribed circles with the midpoints of the four median lines as the centers, and count the number of abnormal points and the total number of pixels in the four pixel circles; The pixel circle whose number of abnormal points is less than four-fifths of the total number of pixel points is recorded as a normal circle, and the number of all normal circles in A building remote sensing images is counted to obtain A normal value; The radius of the pixel circle, the length and width of the building remote sensing image are measured by the scale, and the area of ​​the normal circle is calculated based on the circle area formula. The areas of all normal circles are added together and compared with the total area of ​​the building remote sensing image to obtain A regional completeness. The expression of regional integrity is: ； In the formula, For the The regional completeness of the building remote sensing image, For the The normal value of a building remote sensing image, is the radius of the pixel circle, For the The length of a building remote sensing image, For the The width of a building remote sensing image.

4. A database management system based on building remote sensing images according to claim 3, characterized in that: The target image recognition method is: When the abnormal area ratio of the building remote sensing image is less than the abnormal area safety value, and the area integrity is greater than the area integrity safety value, the building remote sensing image is recorded as the target image, and D target images are obtained; When the abnormal area ratio is greater than or equal to the abnormal area safety value, or the area integrity is less than or equal to the area integrity safety value, the building remote sensing image will not be recorded as the target image.

5. A database management system based on building remote sensing images according to claim 4, characterized in that: The method for extracting the contour color is: The buildings in the D target images are identified one by one by computer technology, the edge lines of the buildings are identified by edge detection algorithms, and the area inside the edge lines is recorded as the building area; Mark the pixel values ​​of all the pixels in the building area on the D target images respectively, and aggregate the pixels with the same color level to obtain E pixel sets; The number of pixels in the E pixel sets is counted one by one, and the color level with the maximum number of pixels is recorded as D valid levels; The natural language processing technology is used to identify D effective levels of text semantics and digital semantics, and after combining the text semantics and digital semantics head to tail, D contour colors are generated.

6. A database management system based on building remote sensing images according to claim 5, characterized in that: The image set includes a semi-hidden set and an open set. The division method of the semi-hidden set and the open set is: According to the order of image time, the D target images are numbered in ascending order, with 1 being the first number; After the target images with consistent geographic coordinates are aggregated in ascending order of numbering, a first image group is generated; Among the target images with inconsistent geographic coordinates, target images with consistent outline colors and building structures are aggregated to generate a second image group; After all the target images in the first image group and the second image group are combined, an open set is generated, and after all the remaining target images are summarized, a semi-hidden set is generated.

7. A database management system based on building remote sensing images according to claim 6, characterized in that: The image database is generated as follows: A blank data group with a closed structure is established, and two internal and external combined data layers are constructed in the blank data group, and the internal data layer is recorded as the internal layer, and the external data layer is recorded as the external layer; Count the number of target images in the semi-hidden set and the open set respectively to obtain the hidden value and the open value; Mark the external image bits in the external layer in the same number as the open value, and mark the internal image bits in the internal layer in the same number as the hidden value, to construct the library architecture; Import all target images in the semi-hidden set into the inner image position of the inner layer one by one, forcing the inner layer to be converted into a hidden layer; All target images in the open set are imported into the external image position of the external layer one by one, forcing the external layer to be converted into an open layer, and the library architecture with hidden layers and open layers is recorded as an image database.

8. A database management system based on building remote sensing images according to claim 7, characterized in that: The method of building the associated channel is: A point is randomly marked on the hidden layer and the open layer of the image database, respectively, and recorded as the inner connection point and the outer connection point; Measure the distance between the inner connection point and the outer connection point to obtain the inner and outer connection value; The position of the external connection point on the open layer is continuously adjusted until the internal and external connection values ​​reach the minimum value, and the position of the external connection point is stopped, and one quarter of the minimum value of the internal and external connection value is recorded as the interval value; An associated channel with a bidirectional transmission link is built between the inner connection point and the adjusted outer connection point, and a point corresponding to a length of an interval value from the outer connection point is marked on the associated channel and recorded as a connection point.

9. A database management system based on building remote sensing images according to claim 8, characterized in that: The on / off state includes the on state and the off state: The on-off condition is: when the open layer is not sufficient to meet the query requirements, the control on-off point is switched from the closed state to the open state.

10. A database management system based on building remote sensing images according to claim 9, characterized in that: The on / off status identification method is: Identify the request semantics in the request data through natural speech processing technology, and extract the keywords in the request semantics one by one; Time, coordinates, color and structure are recorded as demand words, keywords in the request data containing any demand word are recorded as valid words, and all valid words are combined to generate query requirements; Count the number of valid words in the query requirements and record it as the effective value; When the effective value is 1 and the effective word is coordinate, the control on-off point is in the closed state; When the effective value is 1 and the effective word is not a coordinate, the control on-off point is in the on state; When the effective value is 2, 3 or 4, the control on / off point is in the open state; The query mode includes independent query mode and combined query mode. The selection method of independent query mode and combined query mode is as follows: When the on / off state of the on / off point is on, select the combined query mode; When the on / off state of the on / off point is off, select the independent query mode.

Images