CN115292530A - Remote sensing image overall management system - Google Patents
Remote sensing image overall management system Download PDFInfo
- Publication number
- CN115292530A CN115292530A CN202211205144.1A CN202211205144A CN115292530A CN 115292530 A CN115292530 A CN 115292530A CN 202211205144 A CN202211205144 A CN 202211205144A CN 115292530 A CN115292530 A CN 115292530A
- Authority
- CN
- China
- Prior art keywords
- module
- image
- remote sensing
- retrieval
- sensing image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000002372 labelling Methods 0.000 claims abstract description 28
- 238000007726 management method Methods 0.000 claims abstract description 24
- 238000007689 inspection Methods 0.000 claims abstract description 18
- 238000013500 data storage Methods 0.000 claims abstract description 17
- 238000011156 evaluation Methods 0.000 claims description 43
- 230000007246 mechanism Effects 0.000 claims description 10
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000010801 machine learning Methods 0.000 claims description 4
- 238000013441 quality evaluation Methods 0.000 claims description 4
- 230000010354 integration Effects 0.000 claims description 3
- 238000000034 method Methods 0.000 abstract description 16
- 230000008569 process Effects 0.000 abstract description 9
- 206010063385 Intellectualisation Diseases 0.000 abstract 1
- 230000006870 function Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000004801 process automation Methods 0.000 description 1
- 238000011158 quantitative evaluation Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/51—Indexing; Data structures therefor; Storage structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/53—Querying
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/53—Querying
- G06F16/538—Presentation of query results
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/70—Labelling scene content, e.g. deriving syntactic or semantic representations
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Remote Sensing (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Astronomy & Astrophysics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a remote sensing image overall planning management system, which belongs to the field of remote sensing image overall planning and comprises a remote sensing data acquisition module, a labeling module, a quality inspection module, a data storage module, a retrieval module, a user feedback module, a recommendation module and an output module; the labeling module and the quality inspection module are used for adding a label table and quality inspection score information to the image, so that the image can be intelligently understood before being put into a warehouse; intelligently understanding retrieval information of a user through a retrieval module and carrying out optimal strategy image retrieval based on an algorithm to obtain an image group meeting the user requirement; the user feedback module and the recommendation module are used for secondarily understanding the user requirements and updating the retrieval strategy, so that accurate image recommendation is realized; the system can realize a whole set of image overall planning process which is high in intellectualization, automation and image understanding degree and meets the user requirements, and greatly reduces the threshold of using the system.
Description
Technical Field
The invention belongs to the field of remote sensing image overall planning, and particularly relates to a remote sensing image overall planning management system.
Background
With the increasing accumulation of high-resolution remote sensing satellite data resources, remote sensing applications in various fields are gradually expanded, and the method plays a role in important engineering management. How to store, manage and call mass remote sensing data more orderly and efficiently and how to realize rapid sharing and application of the remote sensing information becomes one of the key concerns in the field of spatial information science, remote sensing data service application mechanisms and the like.
The existing remote sensing image service platform and management system have the following defects: the intelligent degree is not enough, the comprehension degree of the overall system to the images is low, and a user needs higher professional knowledge storage when using the system, so that the system is not friendly to general users; the automation degree is low, and the overall planning meeting the user requirements cannot be realized in full process automation, so that the problem of manpower consumption still exists.
Disclosure of Invention
In order to solve the technical problems, the invention provides a remote sensing image overall management system which can provide an overall management service with high intelligence, high image understanding degree and high automation.
In order to achieve the technical purpose, the technical scheme of the invention is as follows:
the invention provides a remote sensing image overall management system, which comprises:
the remote sensing data acquisition module is used for receiving remote sensing images;
the labeling module is connected with the remote sensing data acquisition module and used for labeling the remote sensing image and generating a label table;
the quality inspection module is connected with the labeling module, receives the remote sensing image output by the labeling module and is used for evaluating the content quality of the remote sensing image to obtain an evaluation score;
the data storage module is connected with the quality inspection module and used for storing the remote sensing image output by the quality inspection module and the corresponding label table and evaluation score;
the retrieval module is connected with the data storage module and used for receiving the retrieval information, generating a retrieval tag and searching the remote sensing image stored in the data storage module according to the retrieval tag to obtain an image group;
the output module is connected with the retrieval module and used for receiving the image group output by the retrieval module and outputting the image group to a user in a thematic map form;
the user feedback module is connected with the output module and used for receiving feedback information of the user and adding the feedback information of the user to the output image group;
and the recommendation module is respectively connected with the user feedback module and the retrieval module and used for modifying the retrieval tag according to the feedback information of the user and transferring the modified retrieval tag to the retrieval module for re-searching the remote sensing image.
In an embodiment of the present invention, the labeling module includes:
the ground object labeling module is used for obtaining a ground object label of the remote sensing image according to a machine learning algorithm;
the scene labeling module is used for generating a scene label of the remote sensing image;
and the label integration module is used for integrating the ground feature label and the scene label to generate a label table.
In an embodiment of the present invention, the scene labeling module includes:
the scene classification processing unit is used for carrying out scene classification on the remote sensing image to obtain a scene data set;
an encoder for extracting global features, local features and semantic descriptions corresponding to each feature of a scene data set;
the attention mechanism is used for paying attention to semantic description to different degrees to obtain image features;
and the decoder is used for decoding the image characteristics to generate natural sentence description of the scene data set, namely the scene label of the remote sensing image.
In an embodiment of the present invention, "paying attention to semantic descriptions to different degrees by using an attention mechanism to obtain image features" includes:
calculating attention weight of the local features by adopting an attention weight calculation formula based on the global features;
obtaining image features based on the local features and attention weights thereof;
wherein, the attention weight calculation formula is as follows:
in order to take care of the weight of attention,andis the size of the weight, and,is a global feature of the remote sensing image,is a local feature of the remote sensing image.
In an embodiment of the present invention, the quality inspection module includes:
the quality evaluation module is used for determining an evaluation item of the content quality of the remote sensing image and evaluating the content quality of the remote sensing image based on the evaluation item to obtain evaluation information;
and the score conversion module is used for converting the evaluation information into a score to obtain an evaluation score.
In an embodiment of the present invention, the data storage module includes:
the image storage module is used for storing the remote sensing image;
the tag table storage module is used for storing a tag table;
and the evaluation information storage module is used for storing the evaluation scores.
In an embodiment of the present invention, the retrieving module includes:
the user input module is used for receiving retrieval information input by a user, wherein the retrieval information is multi-modal and comprises retrieval conditions, semantic information and sample drawings;
the retrieval information conversion module is used for converting the retrieval information into label information and combining the label information to obtain a retrieval label;
and the image searching module searches the images stored by the image storage module based on the retrieval tag and the searching algorithm and outputs an image group.
In an embodiment of the present invention, the search condition includes an image resolution, an image star source, a sensor, an image range, an image time, and an image product grade;
the semantic information includes text information and voice information.
In a specific embodiment of the present invention, the feedback information of the user includes images that the user selects to satisfy the user requirement and images that do not satisfy the user requirement in the image group by clicking.
The beneficial effects of the invention are as follows: the invention provides a remote sensing image overall management system which can provide overall management service with high intelligence, high image understanding degree and high automation and provides a lower use threshold for users with low professional knowledge degree.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required in the embodiments will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a block diagram illustrating an embodiment of a system for overall management of remote sensing images according to the present invention;
FIG. 2 is a flowchart illustrating a remote sensing image overall management system according to an embodiment of the present invention;
FIG. 3 is a flowchart illustrating the work flow of image warehousing in an embodiment of the overall remote sensing image management system according to the present invention;
FIG. 4 is a flowchart illustrating image retrieval according to an embodiment of the present invention;
fig. 5 is a flowchart illustrating image recommendation in an embodiment of the overall remote sensing image management system according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments that can be derived by one of ordinary skill in the art from the embodiments given herein are intended to be within the scope of the present invention.
Referring to fig. 1, fig. 1 is a block diagram of an embodiment of a remote sensing image overall planning management system according to the present invention, where the system includes a remote sensing data acquisition module, a labeling module, a quality inspection module, a database module, a retrieval module, a user feedback module, a recommendation module, and an output module.
The overall work flow of the system is shown in fig. 2, and the specific module connections and functions are as follows:
and the remote sensing data acquisition module is used for receiving the remote sensing image. The remote sensing image can be image data of original satellite data which is processed by data analysis, uniform radiation correction, denoising, MTFC, CCD splicing and wave band registration.
And the marking module is connected with the remote sensing data acquisition module and is used for marking the remote sensing image and generating a label table.
The annotation module can include:
the ground object labeling module is used for obtaining a ground object label of the remote sensing image according to a machine learning algorithm;
the scene labeling module is used for generating a scene label of the remote sensing image;
and the label integration module is used for integrating the ground feature label and the scene label to generate a label table. The tag table may be constructed by using the feature tag as the first column of data and the scene tag as the second column of data.
Specifically, the scene labeling module includes:
the scene classification processing unit is used for carrying out scene classification on the remote sensing image to obtain a scene data set;
an encoder for extracting global features, local features and semantic descriptions corresponding to each feature of a scene data set;
the attention mechanism is used for paying attention to semantic description to different degrees to obtain image features;
and the decoder is used for decoding the image characteristics to generate natural sentence description of the scene data set, namely the scene label of the remote sensing image.
Specifically, a specific implementation manner of obtaining the scene tag by using the scene labeling module is as follows:
and carrying out scene classification on the remote sensing image by using a scene classification processing unit to obtain a scene data set. The scene classification processing unit mainly comprises a classification prediction layer, a cross entropy loss function and an optimizer. During training, a scene classification result of the remote sensing sample data is obtained through forward propagation, a cross entropy loss function is calculated, a momentum optimizer is selected, and model parameters are adjusted through iterative training. And carrying out scene classification on the remote sensing image by using the trained scene classification processing unit.
Global features, local features and a semantic description corresponding to each feature of the scene data set are then extracted based on the encoder. The encoder may be constructed by a deep convolutional network, and the embodiment adopts the VGG16 as the encoder. The process of extracting features is as follows:
the remote sensing image is expressed by S, the top layer features, namely the local features, are extracted by using the middle layer of VGG16, the global features are extracted by using the high layer, and the formula is as follows:
wherein, the first and the second end of the pipe are connected with each other,is a global feature of the remote sensing image,is a local feature of the remote sensing image,for the subsequent attention mechanism operation, the system is used,and generating semantic description of the remote sensing image.
And (4) paying attention to the semantic description to different degrees by using an attention mechanism to obtain image features. The attention mechanism can assign weights with different sizes to different positions in the image, and the importance of each position can be automatically adjusted. The calculation of the attention weights is based on the hidden state vector output by the decoder at the previous moment. Let the output of the decoder LSTM at the last moment beLet us order
Wherein, the first and the second end of the pipe are connected with each other,for attention weighting, the importance of the various regions of the remote sensing image features can be adjusted,andis the size of the weight(s),for performing a dimension transformation ofIs transformed to andthe sizes are consistent.
After the attention weight is calculated, the intermediate layer features, namely the local features, are subjected to weighted summation.
Wherein the content of the first and second substances,the context feature vector of the image features comprises overall information of each region of the remote sensing image after importance adjustment, j represents the jth image region on the image feature map, and N represents the number of the regions contained in the remote sensing image.
And decoding the image characteristics based on a decoder to generate natural sentence description of the scene data set, namely a scene label of the remote sensing image. The decoder can be built by utilizing a cyclic recursive network, after the context characteristic vector is obtained, word embedding is carried out on the word through an embedding layer, the word embedding vector of the word and the context characteristic vector are spliced, the output of the decoder at the current moment is obtained, the hidden state at the next moment can be predicted through the decoder according to the hidden state vector at the previous moment, then the hidden state is mapped into a space with the same dimension as the dictionary through a transformation matrix, and the probability distribution of the word output at the next moment can be predicted through a SoftMax function. And selecting the word corresponding to the maximum probability as output by adopting a greedy algorithm according to the distribution of the word prediction probability, thereby obtaining the natural sentence description of the remote sensing image, namely the scene label.
And the quality inspection module is connected with the labeling module, receives the image output by the labeling module and is used for evaluating the content quality of the remote sensing image to obtain an evaluation score.
Further, the quality inspection module comprises:
the quality evaluation module is used for determining an evaluation item of the content quality of the remote sensing image and evaluating the content quality of the remote sensing image based on the evaluation item to obtain evaluation information; the evaluation items comprise definition evaluation, mean square error evaluation, information entropy evaluation, peak signal-to-noise ratio evaluation, invalid pixel evaluation, multispectral and panchromatic consistency evaluation, wave band matching evaluation, side-view angle overrun evaluation, cloud cover evaluation, shadow evaluation, strip evaluation, boundary abnormity evaluation and histogram abnormity evaluation;
and the score conversion module is used for converting the evaluation information into a score to obtain an evaluation score.
And the data storage module is connected with the quality inspection module and is used for storing the image output by the quality inspection module and the label table and the evaluation information corresponding to the image.
Further, the data storage module includes:
the image storage module is used for storing image data;
the tag table storage module is used for storing a tag table;
and the evaluation information storage module is used for storing the evaluation scores.
The process that the original image reaches the data storage module through the remote sensing data acquisition module, the labeling module and the quality inspection module belongs to the image warehousing process, and is shown in fig. 3.
And the retrieval module is connected with the data storage module and used for receiving the retrieval information, generating a retrieval tag and searching the remote sensing image stored in the data storage module according to the retrieval tag to obtain an image group.
Further, the retrieval module includes:
the user input module is used for receiving retrieval information input by a user; the retrieval information is multi-modal and comprises retrieval conditions, semantic information and sample graphs, and the user input module can simultaneously receive at least one type of retrieval information. The retrieval conditions are inherent options provided by the system, including image resolution, image star source, sensor, image range, image time and image product grade, and the user selects the options in a manual click mode. The semantic information includes text information and voice information. The user provides the retrieval information in the form of input text or voice, and the voice information is converted into text information after voice recognition.
And the retrieval information conversion module is used for converting the multi-mode retrieval information into label information and combining the label information to obtain a retrieval label. Wherein the module converts semantic information into tag information by performing semantic understanding on the semantic information. The semantic understanding method can be used for segmenting semantic information, labeling keywords of the semantic information according to words obtained by segmenting and a preset knowledge base, and matching the keywords with the image tag types to generate a semantic understanding result. For example, the semantic information input by the user is "i need zheng state resolution two meters of images", the keywords obtained by keyword labeling include "zheng state" and "2m", the zheng state "belongs to the administrative division label, and the" 2m "belongs to the resolution label, and the keywords are used as the retrieval label. Wherein the module converts the sample graph into label information by performing image understanding on the sample graph. The image understanding method can be used for extracting features of the sample graph and converting the extracted features into label information.
And the image searching module searches the images stored by the image storage module based on the retrieval tag and the image searching algorithm and outputs an image group, the image searching algorithm can comprise an image retrieval algorithm based on texture features, an image retrieval algorithm based on shape features, an algorithm based on a Hash algorithm, an algorithm based on an A-Star algorithm, an algorithm based on a Monte Carlo tree algorithm and an algorithm based on an evolutionary algorithm, and the path selected by the images is determined through the algorithms.
The output module is connected with the retrieval module and used for receiving the image group data output by the retrieval module and outputting the image group data to a user in a thematic map form;
the user inputs the search information, and the search module performs the image search process in the data storage module according to the search information, as shown in fig. 4.
The user feedback module is connected with the output module and used for receiving feedback information of the user and adding the feedback information to the output image group, and the user can select images meeting user requirements and images not meeting the user requirements in the retrieval result in a click mode to perform feedback;
and the recommendation module is connected with the user feedback module and the retrieval module and used for modifying the retrieval tag according to the user feedback information and transferring the modified user retrieval tag to the retrieval module for re-searching the remote sensing image. The method for modifying the retrieval tag can be to externally expand the retrieval tag and appropriately expand the search range. The search tags capable of being externally extended can be image time, sensor, image quality and product grade.
The user inputs the feedback information, and the recommendation module modifies the retrieval information according to the feedback information and performs a workflow of image recommendation, as shown in fig. 5.
The working principle and the using process of the invention are as follows: when a user starts to use the remote sensing image overall management system, the remote sensing image is received by the remote sensing data acquisition module, the remote sensing image is intelligently marked by the marking module, and the ground object label and the scene label of the remote sensing image are generated by the ground object marking module using a machine learning algorithm and the scene marking module using an attention mechanism, so that a label table is further formed, and the comprehension capability of the system to the remote sensing image is improved. The quality inspection module carries out further quantitative evaluation on the remote sensing image, gives a comprehensive quality score for the remote sensing image based on various quality evaluation items, and facilitates the system to carry out subsequent remote sensing image retrieval and recommendation according to the quality of the remote sensing image. The data storage module receives the processed remote sensing image, and stores the remote sensing image and the corresponding label table and evaluation score in a standard manner, so that the remote sensing image can be conveniently selected directly through the label table and the evaluation score during subsequent retrieval and recommendation. The retrieval module receives retrieval information of a user, intelligently understands the retrieval information, obtains the remote sensing image meeting the user requirement by matching the retrieval information with a tag table of the remote sensing image in the data storage module, maximizes the total mass fraction of the image group by a retrieval algorithm and realizes an optimal image group search strategy. After a group of image search results are obtained, a user can also feed back the search results through the user feedback module, the recommendation module modifies the search information based on understanding according to the user feedback information and sends the modified search information to the search module, the search module carries out image search again and outputs the search results to the user, the user can feed back the search results for multiple times until the search results completely meet the user requirements, and the accuracy of image search and recommendation is further guaranteed.
The invention has the beneficial effects that: the invention provides a remote sensing image overall management system which can provide overall management service with high intelligence, high image understanding degree and high automation and provides a lower use threshold for users with low professional knowledge degree.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising a … …" does not exclude the presence of another identical element in a process, method, article, or apparatus that comprises the element.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (9)
1. The utility model provides a remote sensing image overall management system which characterized in that includes:
the remote sensing data acquisition module is used for receiving remote sensing images;
the labeling module is connected with the remote sensing data acquisition module and used for labeling the remote sensing image and generating a label table;
the quality inspection module is connected with the labeling module, receives the remote sensing image output by the labeling module and is used for evaluating the content quality of the remote sensing image to obtain an evaluation score;
the data storage module is connected with the quality inspection module and is used for storing the remote sensing image output by the quality inspection module and the corresponding label table and evaluation score;
the retrieval module is connected with the data storage module and used for receiving the retrieval information, generating a retrieval tag and searching the remote sensing image stored in the data storage module according to the retrieval tag to obtain an image group;
the output module is connected with the retrieval module and used for receiving the image group output by the retrieval module and outputting the image group to a user in a thematic map form;
the user feedback module is connected with the output module and used for receiving feedback information of the user and adding the feedback information of the user to the output image group;
and the recommendation module is respectively connected with the user feedback module and the retrieval module and used for modifying the retrieval tag according to the feedback information of the user and transferring the modified retrieval tag to the retrieval module for re-searching the remote sensing image.
2. The system as claimed in claim 1, wherein the labeling module comprises:
the ground object labeling module is used for obtaining a ground object label of the remote sensing image according to a machine learning algorithm;
the scene labeling module is used for generating a scene label of the remote sensing image;
and the label integration module is used for integrating the ground feature label and the scene label to generate a label table.
3. The system as claimed in claim 2, wherein the scene labeling module comprises:
the scene classification processing unit is used for carrying out scene classification on the remote sensing image to obtain a scene data set;
an encoder for extracting global features, local features and semantic descriptions corresponding to each feature of a scene data set;
the attention mechanism is used for paying attention to semantic description to different degrees to obtain image features;
and the decoder is used for decoding the image characteristics to generate natural sentence description of the scene data set, namely the scene label of the remote sensing image.
4. The remote sensing image overall planning management system according to claim 3, wherein the step of paying attention to semantic descriptions to different degrees by using an attention mechanism to obtain image features comprises the steps of:
calculating attention weight of the local features by adopting an attention weight calculation formula based on the global features;
obtaining image features based on the local features and the attention weights thereof;
wherein, the attention weight calculation formula is as follows:
5. The system as recited in claim 1, wherein the quality inspection module comprises:
the quality evaluation module is used for determining an evaluation item of the content quality of the remote sensing image and evaluating the content quality of the remote sensing image based on the evaluation item to obtain evaluation information;
and the score conversion module is used for converting the evaluation information into scores to obtain evaluation scores.
6. The remote sensing image orchestration management system according to claim 1, wherein the data storage module comprises:
the image storage module is used for storing the remote sensing image;
the tag table storage module is used for storing a tag table;
and the evaluation information storage module is used for storing the evaluation scores.
7. The remote sensing image orchestration management system according to claim 1, wherein the retrieval module comprises:
the user input module is used for receiving retrieval information input by a user, wherein the retrieval information is multi-modal and comprises retrieval conditions, semantic information and sample drawings;
the retrieval information conversion module is used for converting the retrieval information into label information and combining the label information to obtain a retrieval label;
and the image searching module searches the images stored by the image storage module based on the retrieval tag and the searching algorithm and outputs an image group.
8. The remote sensing image overall planning management system according to claim 7, wherein the retrieval conditions include image resolution, image star source, sensor, image range, image time, image product grade;
the semantic information includes text information and voice information.
9. The remote sensing image overall planning management system of claim 1, wherein the feedback information of the user includes images that the user selects to meet the user's needs and images that do not meet the user's needs from the image group by clicking.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211205144.1A CN115292530A (en) | 2022-09-30 | 2022-09-30 | Remote sensing image overall management system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211205144.1A CN115292530A (en) | 2022-09-30 | 2022-09-30 | Remote sensing image overall management system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115292530A true CN115292530A (en) | 2022-11-04 |
Family
ID=83833998
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211205144.1A Pending CN115292530A (en) | 2022-09-30 | 2022-09-30 | Remote sensing image overall management system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115292530A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117098282A (en) * | 2023-10-18 | 2023-11-21 | 硬创(济宁)控股有限公司 | Big data management system and method based on intelligent park |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106909924A (en) * | 2017-02-18 | 2017-06-30 | 北京工业大学 | A kind of remote sensing image method for quickly retrieving based on depth conspicuousness |
CN111860235A (en) * | 2020-07-06 | 2020-10-30 | 中国科学院空天信息创新研究院 | Method and system for generating high-low-level feature fused attention remote sensing image description |
US20200372289A1 (en) * | 2019-05-22 | 2020-11-26 | Here Global B.V. | Method, apparatus, and system for priority ranking of satellite images |
CN112232151A (en) * | 2020-09-30 | 2021-01-15 | 河海大学 | Iterative aggregation neural network high-resolution remote sensing scene classification method embedded with attention mechanism |
CN115080774A (en) * | 2022-07-20 | 2022-09-20 | 北京数慧时空信息技术有限公司 | Remote sensing image warehousing system and method based on available domain |
-
2022
- 2022-09-30 CN CN202211205144.1A patent/CN115292530A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106909924A (en) * | 2017-02-18 | 2017-06-30 | 北京工业大学 | A kind of remote sensing image method for quickly retrieving based on depth conspicuousness |
US20200372289A1 (en) * | 2019-05-22 | 2020-11-26 | Here Global B.V. | Method, apparatus, and system for priority ranking of satellite images |
CN111860235A (en) * | 2020-07-06 | 2020-10-30 | 中国科学院空天信息创新研究院 | Method and system for generating high-low-level feature fused attention remote sensing image description |
CN112232151A (en) * | 2020-09-30 | 2021-01-15 | 河海大学 | Iterative aggregation neural network high-resolution remote sensing scene classification method embedded with attention mechanism |
CN115080774A (en) * | 2022-07-20 | 2022-09-20 | 北京数慧时空信息技术有限公司 | Remote sensing image warehousing system and method based on available domain |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117098282A (en) * | 2023-10-18 | 2023-11-21 | 硬创(济宁)控股有限公司 | Big data management system and method based on intelligent park |
CN117098282B (en) * | 2023-10-18 | 2024-01-09 | 硬创(济宁)控股有限公司 | Big data management system and method based on intelligent park |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11699298B2 (en) | Training method of image-text matching model, bi-directional search method, and relevant apparatus | |
CN110795590B (en) | Multi-label image retrieval method and device based on direct-push zero-sample hash | |
CN111581510A (en) | Shared content processing method and device, computer equipment and storage medium | |
CN107368614A (en) | Image search method and device based on deep learning | |
CN110795527B (en) | Candidate entity ordering method, training method and related device | |
CN111666427A (en) | Entity relationship joint extraction method, device, equipment and medium | |
Rad et al. | Image annotation using multi-view non-negative matrix factorization with different number of basis vectors | |
CN110688474B (en) | Embedded representation obtaining and citation recommending method based on deep learning and link prediction | |
CN111694965A (en) | Image scene retrieval system and method based on multi-mode knowledge graph | |
CN114186084B (en) | Online multi-mode Hash retrieval method, system, storage medium and equipment | |
CN113051914A (en) | Enterprise hidden label extraction method and device based on multi-feature dynamic portrait | |
CN113806580B (en) | Cross-modal hash retrieval method based on hierarchical semantic structure | |
CN111738169A (en) | Handwriting formula recognition method based on end-to-end network model | |
CN111985520A (en) | Multi-mode classification method based on graph convolution neural network | |
CN113377981A (en) | Large-scale logistics commodity image retrieval method based on multitask deep hash learning | |
CN112948601A (en) | Cross-modal Hash retrieval method based on controlled semantic embedding | |
CN111241310A (en) | Deep cross-modal Hash retrieval method, equipment and medium | |
CN115292530A (en) | Remote sensing image overall management system | |
CN115357747A (en) | Ordinal hash-based image retrieval method and system | |
CN113642602A (en) | Multi-label image classification method based on global and local label relation | |
CN115878757A (en) | Concept decomposition-based hybrid hypergraph regularization semi-supervised cross-modal hashing method | |
CN113516118B (en) | Multi-mode cultural resource processing method for joint embedding of images and texts | |
CN113139110B (en) | Regional characteristic processing method, regional characteristic processing device, regional characteristic processing equipment, storage medium and program product | |
CN112925934B (en) | Similar image retrieval method, system, device and medium based on Hash coding | |
CN115063831A (en) | High-performance pedestrian retrieval and re-identification method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20221104 |
|
RJ01 | Rejection of invention patent application after publication |