Remote sensing image analysis method and device

Abstract

The invention provides a remote sensing image analysis method and device, and belongs to the technical field of remote sensing image processing. The method comprises the following steps: receiving multi-source multi-temporal remote sensing image data, and constructing a space-time spectrum multi-dimensional data cube based on a space grid engine; target grid data is dynamically scheduled according to a user request, real-time processing such as dynamic stretching, wave band combination and light and color uniformity is carried out, and the target grid data is published in a map service form. The device comprises a data receiving module, a scheduling module, a dynamic processing module and a service publishing module. Through gridding scheduling, dynamic processing and micro-service architecture, the slicing-free, multi-specification and high-concurrency online service of the remote sensing image is realized, the data processing efficiency, the storage utilization rate and the service stability are remarkably improved, and the method is suitable for real-time analysis and application of mass remote sensing data.

Description

A method and apparatus for remote sensing image analysis Technical Field

[0001] This invention relates to the field of remote sensing image analysis technology, and in particular to a remote sensing image analysis method and apparatus. Background Technology

[0002] In remote sensing image analysis, traditional analysis methods rely on preprocessing and offline analysis, which cannot meet the needs of real-time analysis and online services, especially when dealing with massive amounts of remote sensing images from multiple time periods and sources.

[0003] Therefore, there is an urgent need in this field for an analysis method that can provide real-time analysis and online services to improve the application efficiency of remote sensing images.

[0004] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide an analysis method that enables real-time analysis and online services.

[0006] To achieve the above objectives, the present invention provides the following solution:

[0007] A remote sensing image analysis method includes the following steps: receiving multi-source, multi-temporal remote sensing image data; performing logical discretization processing on the remote sensing image data based on a spatial grid engine to construct a spatiotemporal spectral multidimensional data cube; dynamically scheduling target grid data in the data cube according to user requests; performing real-time dynamic processing on the target grid data, including at least one of dynamic stretching enhancement, dynamic band combination, dynamic slicing, and dynamic light and color homogenization; and publishing the processed data in real-time as a map service.

[0008] Optionally, the spatial grid engine adopts a global subdivision grid coding system, which maps spatial locations to unique grid codes to achieve pixel-level data scheduling.

[0009] Optionally, in the dynamic processing step, users can adjust the stretching parameters and band combination order in real time through the front-end interface, and the result is rendered and displayed in the browser in real time.

[0010] Optional features include: performing time sorting and dynamic mosaicking of multi-temporal remote sensing image data to generate a unified WMTS service; and supporting dynamic filtering and display of image data from different time phases based on time parameters.

[0011] Optional features include: visualizing and compressing the original remote sensing image data with low loss, and supporting dynamic service publishing based on the compressed data.

[0012] Optionally, the method adopts a microservice architecture, supports multi-node cluster deployment and load balancing, and achieves stable service under high concurrency access.

[0013] A remote sensing image analysis device includes: a data receiving module for receiving multi-source, multi-temporal remote sensing image data; a data scheduling module for logically discretizing and scheduling the data based on a spatial grid engine; a dynamic processing module for real-time dynamic processing of the scheduled large dataset; and a service publishing module for publishing the processing results in the form of a standard map service.

[0014] Optionally, the dynamic processing module supports dynamic stretching enhancement, dynamic band combination, dynamic light and color homogenization, and dynamic data fusion functions.

[0015] Optionally, the service publishing module supports the WMTS service protocol and allows for dynamic data filtering based on time, space, and band parameters.

[0016] Optionally, the device is deployed in a distributed cluster environment, supporting horizontal scaling and load balancing, and possessing high availability and high concurrency processing capabilities.

[0017] Compared with existing technologies, this invention has the following advantages: It provides a remote sensing image analysis method and apparatus that achieves pixel-level data scheduling through a spatial grid engine, supporting dynamic stretching, band combination, and uniform illumination and color processing. Users can adjust parameters and preview effects in real time without pre-tiling. The use of logical discretization grid scheduling technology reduces data redundancy and supports visual low-loss compression (compressed to 4% of the original file size), significantly saving storage space. Multi-temporal images are published as a single WMTS service, supporting dynamic filtering and backtracking in the time dimension, avoiding multi-layer call bottlenecks and improving front-end rendering efficiency. The system supports distributed cluster deployment, load balancing, and horizontal scaling, possessing high availability and high concurrency processing capabilities. One set of data supports multiple map service specifications, eliminating the need for repeated tiling and improving data utilization and publishing efficiency. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 is a schematic diagram of the remote sensing image analysis method provided in an embodiment of the present invention.

[0020] Figure 2 is a schematic diagram of the remote sensing image analysis device provided in an embodiment of the present invention.

[0021] Figure 3 is a schematic diagram of the parameter-driven multi-specification dynamic map service provided in an embodiment of the present invention.

[0022] Figure 4 is a schematic diagram of the slice-free publishing provided by an embodiment of the present invention.

[0023] Figure 5 is a schematic diagram of API gateway and microservice management provided in an embodiment of the present invention.

[0024] Figure 6 is a schematic diagram of the cluster service architecture provided in an embodiment of the present invention.

[0025] Figure 7 is a structural diagram of the data organization model based on spatial grid provided in an embodiment of the present invention.

[0026] Figure 8 is a schematic diagram of the parameter-driven, tile-free, multi-specification dynamic map service mode provided in an embodiment of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] The purpose of this invention is to provide a method and apparatus for remote sensing image analysis that can provide real-time analysis and online services.

[0029] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0030] Example 1: A remote sensing image analysis method includes the following steps: receiving multi-source, multi-temporal remote sensing image data; performing logical discretization processing on the remote sensing image data based on a spatial grid engine to construct a spatiotemporal spectral multidimensional data cube; dynamically scheduling target grid data in the data cube according to user requests; specifically, dynamic scheduling involves: constructing a logical row-column model of remote sensing data; parsing the physical address of data content storage; and using a file pointer offset method, reading the internal bytes of the file involved in the grid segment by grid unit without loading the entire file, thus realizing grid scheduling technology for remote sensing data. Based on grid-oriented data scheduling technology, the organization model of remote sensing data is changed, and the data reading mode is transformed from file-oriented to grid-oriented, forming a pixel-level data scheduling mode, realizing "content-oriented spatial data structured management".

[0031] The grid rules are dynamic, and the spatial units for data reading are also dynamic. The grid definition can be multi-rule and multi-level. Traditional slicing rules can be set as grid rules, or different grid scheduling rules can be designed for specific applications.

[0032] By reducing the data volume of each data call, routine data processing can be performed in memory in real time; by leveraging the discrete nature of the grid, multiple processes can be started simultaneously for parallel reading, improving reading speed.

[0033] The target grid data is processed dynamically in real time, including at least one of dynamic stretching enhancement, dynamic band combination, and dynamic light and color equalization. Remote sensing image dynamic tiling, dynamic enhancement, and band combination technologies are employed: tiling rules are understood as grid rules, tiling algorithms are managed, and a grid data scheduling method is used to acquire data from one tile range at a time. Data cropping, merging, thinning, and compression are performed in memory to achieve dynamic tiling. Due to the dynamic processing, different tile rules can be used for dynamic tiling, successfully realizing a "parameter-driven multi-specification map service," as shown in Figure 3. Multiple tiling rules are built-in and extensible; different specifications of map services can be obtained by calling different parameters from the application front-end. This changes the traditional model where different rules require generating multiple sets of tiling data and publishing them to different service addresses, avoiding inefficiency, repetitive tiling of multiple specifications, and difficulties in dynamic updates. Only one image library is needed, eliminating the need to generate massive amounts of tiling data, significantly improving the efficiency of remote sensing data applications.

[0034] Thanks to the use of dynamic processing technology, automatic publishing without slicing can be achieved. As shown in Figure 4, the new technical approach of direct publishing without parameters and unified publishing of mixed specifications has been tested and can improve the efficiency of remote sensing data publishing by more than 1,000 times.

[0035] We developed a dynamic stretching and enhancement technology that can dynamically stretch and enhance 4-band 16-bit images and directly display 3-band 8-bit true-color images at the front end. Compared with the traditional mode that requires processing a set of true-color data, this technology reduces processing steps and storage space. We also independently developed a dynamic light and color homogenization technology to unify the color style of different satellite data.

[0036] For single-scene data with unsatisfactory display effects, parameters can be manually adjusted, and only the color adjustment parameters need to be saved to achieve the preset effect during browsing.

[0037] The processed data will be published in real time as a map service.

[0038] The spatial grid engine adopts a global subdivision grid coding system, which maps spatial locations to unique grid codes to achieve pixel-level data scheduling.

[0039] We have independently developed a dynamic band combination technology, which allows us to set the spectral bands and their arrangement order for display in multi-band remote sensing images, and dynamically display them as different false-color effects on the front end. Compared with similar functions in traditional desktop software, this technology does not require data download and can achieve the corresponding functions in a browser.

[0040] In the dynamic processing step, users can adjust the stretching parameters and band combination order in real time through the front-end interface, and the results are rendered and displayed in the browser in real time.

[0041] It also includes: time sorting and dynamic mosaicking of multi-temporal remote sensing image data to generate a unified WMTS service; and supports dynamic filtering and display of image data from different time phases by time parameters.

[0042] It also includes: visualizing and compressing raw remote sensing image data with low loss, and supporting dynamic service publishing based on the compressed data.

[0043] Grid-based data compression technology: With the development of new sensing technologies, the temporal, spatial, and spectral resolution of remote sensing images is constantly improving, and the corresponding data scale is growing exponentially, posing great challenges to data transmission and storage. Data compression technology, as an effective way to solve this problem, is receiving increasing attention in the field of remote sensing. Image compression technology can be divided into two main categories: lossless compression technology and lossy compression technology. Lossless compression utilizes the statistical redundancy of data for compression, which can completely recover the original data without introducing any distortion, but the compression ratio is theoretically limited by the statistical redundancy of data, generally ranging from 2:1 to 5:1. Lossy compression methods utilize the characteristic that human vision is insensitive to certain frequency components in images, allowing for the loss of some information during the compression process; although it cannot completely recover the original data, the lost parts have a smaller impact on understanding the original image, while achieving a much higher compression ratio.

[0044] This project develops dynamic processing technology directly targeting raw spectral data. While offering excellent flexibility, it also places higher demands on IT infrastructure capabilities, including storage, networking, and computing. Limited IT infrastructure can reduce overall system performance. To further optimize operational efficiency, the project team developed a grid-based lossy data compression technique. This technique extracts the visual portions of remote sensing imagery and uses discrete wavelet transform compression technology to perform cross-level data compression based on a large-size grid, reducing file size to less than 4% of the original.

[0045] Through testing, this compression algorithm can complete the compression of a single scene image within one minute using a single server, which fully meets the application requirements for real-time remote sensing image services.

[0046] The generated compressed file also supports dynamic services such as dynamic slicing and dynamic projection. Logically, the compressed file is a cache of the original data; when dynamic band combination display and spectral analysis are needed, the application can still call the original file for processing. The compressed file can also be used independently. In internet environments or for applications requiring only visualization, this compressed file mode can significantly reduce data storage and improve retrieval efficiency.

[0047] Cluster distributed service technology based on microservices: To solve the performance bottleneck of massive data scheduling and dynamic processing, this technology utilizes the logical discretization characteristics of grid data scheduling, combines data scheduling with cloud platforms (distributed storage and distributed computing), and develops a cluster distributed service system using a microservice architecture. This fully leverages the capabilities of IT infrastructure to support the storage, scheduling, and dynamic processing of massive high-resolution remote sensing images, effectively ensuring service performance under massive data and large-scale access.

[0048] As shown in Figure 5, the system adopts a microservice architecture, packaging each backend function into a service component that can be deployed independently. It uses an API gateway for unified management and scheduling, achieving load balancing and service monitoring, and supports container and virtual machine deployment.

[0049] The system supports multi-node cluster deployment. Clustered operation can reduce the number of single points of failure and achieve high availability and high performance of clustered resources. It can also horizontally scale the system's service capacity as data volume and access volume increase.

[0050] As shown in Figure 6, the cluster service consists of a master node and several slave nodes. Load balancing is used to distribute the computing and network pressure brought by service requests. All nodes in the cluster access the same database server and storage resources.

[0051] Data organization model based on spatial grid: as shown in Figure 7; This embodiment provides a set of technical systems for data search, data scheduling, dynamic processing and data services for raw spectral data. It has independently developed a spatial grid engine technology for remote sensing images, which changes the organization model of remote sensing data. A grid layer is added between the application and the data, which isolates the data organization from the data. This changes the traditional data model in which the application directly accesses the data. It makes full use of grid data scheduling technology to form an integrated management and service model for massive remote sensing images.

[0052] Parameter-driven, tile-free, multi-specification dynamic map service mode: As shown in Figure 8, a parameter-driven, tile-free, multi-specification dynamic map service mode was constructed for the first time. It broke through the key technologies of dynamic processing such as stretching enhancement, band combination, and data fusion, and solved the bottlenecks of low efficiency, large storage volume, and difficulty in dynamic updates of traditional process map tiling services. It significantly improved the efficiency of application services and realized a new mode of "one set of data supporting multiple scenarios and multiple types of applications".

[0053] Traditional image management and service technologies rely on tedious, time-consuming, and storage-intensive preprocessing tasks such as image mosaicking, color enhancement, projection transformation, data slicing, and service publishing. When different application scenarios require different image parameter adjustments, re-slicing is necessary, which increases processing time and storage space exponentially, making it impossible to meet the real-time service needs of massive multi-source and multi-temporal remote sensing images.

[0054] This project has established a technical system for data search, data scheduling, dynamic processing, and data services. It replaces the traditional process of "data mosaicking-pre-slicing" with dynamic tiling technology. Through dynamic enhancement, dynamic band combination, and dynamic band fusion technologies, it reduces data preprocessing steps and storage space, significantly saving a lot of manpower and material resources, and enabling "real-time remote sensing image services".

[0055] This project has achieved a management and service model of "one set of data supporting multiple application scenarios". The petabyte-level image library only requires the addition of a small amount of index metadata at the megabyte level, effectively saving more than 40% of storage space. By utilizing the discretization characteristics of grids, a "cloud-native" technical system supporting distributed storage and computing has been constructed.

[0056] A "dynamic single map" service model for massive multi-temporal remote sensing images was constructed, which solved the performance improvement problems caused by multi-temporal and multi-layer service calls and logically discretized data scheduling. It significantly improved the performance of unified scheduling, unified service, fast display and historical backtracking, and effectively improved the stability and reliability of large-scale access to massive remote sensing image data.

[0057] Real-time acquisition and continuous updating of remote sensing data, as well as the ability to monitor long-term time series, are among the core values ​​of remote sensing data. However, the coverage of remote sensing images is unevenly distributed in time and space, and under traditional technical approaches, the management and service of single-scene multi-temporal data faces significant bottlenecks.

[0058] In the traditional model, publishing a single-scene image as an independent service can maintain clear temporal information. Large remote sensing data centers typically manage tens of thousands to hundreds of thousands of images, which means that tens of thousands of layers need to be published. This poses a huge challenge to data management and retrieval. Web front-ends can become inefficient or even crash due to calling too many service layers within a window. When accessing multi-temporal files on the desktop, performance is also low due to layer-by-layer refresh. To solve this problem, traditional techniques often use data mosaicking to form a single layer to avoid the bottleneck of multi-temporal access. This requires a large amount of work and time, resulting in the loss of temporal information and a significant decrease in the utilization rate of remote sensing images.

[0059] This project designed and developed a unified scheduling technology for multi-temporal remote sensing images. While preserving the data organization mode of the original remote sensing image files, it publishes all multi-temporal data under management as a unified service. Through grid search, time sorting, and dynamic mosaicking technology, the content to be displayed is dynamically mosaicked into a WMTS layer according to the time priority principle, so that the front end always calls a single layer, ensuring calling efficiency. This solves the problems of unified scheduling, unified service, and fast display of multi-temporal images. Furthermore, historical data backtracking can be achieved through time dimension filtering, and this capability can be embedded into third-party applications through API calls.

[0060] A remote sensing image analysis device includes: a data receiving module for receiving multi-source, multi-temporal remote sensing image data; a data scheduling module for logically discretizing and scheduling the data based on a spatial grid engine; a dynamic processing module for real-time dynamic processing of the scheduled large data; and supports dynamic stretching enhancement, dynamic band combination, dynamic light and color homogenization, and dynamic data fusion functions.

[0061] The service publishing module is used to publish the processing results in the form of standard map services. It supports the WMTS service protocol and allows dynamic filtering of data by time, space, and band parameters.

[0062] The device is deployed in a distributed cluster environment, supports horizontal scaling and load balancing, and has high availability and high concurrency processing capabilities.

[0063] This embodiment details a remote sensing image analysis method and apparatus. It utilizes a spatial grid engine to achieve logical discretization scheduling and dynamic processing of data, including dynamic tiling, stretching enhancement, and band combination, supporting real-time user interaction. The method employs a microservice architecture, possessing high concurrency processing capabilities, and implements tile-free multi-scale map services through parameter-driven processing. The technical effects are significant: it improves data publishing efficiency by more than a thousand times, supports unified multi-temporal services and historical backtracking, effectively solving problems such as large storage requirements, difficult updates, and low efficiency in multi-layer access in traditional remote sensing image processing, achieving a highly efficient service model of "one set of data for multiple scenarios."

[0064] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0065] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A remote sensing image analysis method, characterized in that, Includes the following steps: Receive multi-source, multi-temporal remote sensing image data; The remote sensing image data is logically discretized based on a spatial grid engine to construct a spatiotemporal spectral multidimensional data cube. Based on user requests, the target grid data in the data cube is dynamically scheduled; the target grid data is processed in real time, including at least one of dynamic stretching enhancement, dynamic band combination, and dynamic light and color equalization; the processed data is published in real time as a map service.

2. The remote sensing image analysis method according to claim 1, characterized in that, The spatial grid engine adopts a global subdivision grid coding system, which maps spatial locations to unique grid codes to achieve pixel-level data scheduling.

3. The remote sensing image analysis method according to claim 1, characterized in that, In the dynamic processing step, users can adjust the stretching parameters and band combination order in real time through the front-end interface, and the results are rendered and displayed in the browser in real time.

4. The remote sensing image analysis method according to claim 1, characterized in that, Also includes: Time-sequencing and dynamic mosaicking of multi-temporal remote sensing image data are performed to generate a unified WMTS service; It supports dynamically filtering and displaying image data at different time phases by using time parameters.

5. The remote sensing image analysis method according to claim 1, characterized in that, Also includes: The system performs low-loss visualization compression on raw remote sensing image data and supports dynamic service publishing based on the compressed data.

6. The remote sensing image analysis method according to claim 1, characterized in that, The method adopts a microservice architecture, supports multi-node cluster deployment and load balancing, and achieves stable service under high concurrency access.

7. A remote sensing image analysis device, characterized in that, include: The data receiving module is used to receive multi-source, multi-temporal remote sensing image data; The data scheduling module is used to perform logical discretization scheduling of data based on the spatial grid engine; the dynamic processing module is used to perform real-time dynamic processing of the scheduled data. The service publishing module is used to publish the processing results in the form of standard map services.

8. The remote sensing image analysis device according to claim 7, characterized in that, The dynamic processing module supports dynamic stretching enhancement, dynamic band combination, dynamic light and color homogenization, and dynamic data fusion functions.

9. The remote sensing image analysis device according to claim 7, characterized in that, The service publishing module supports the WMTS service protocol and allows for dynamic data filtering based on time, space, and band parameters.

10. The remote sensing image analysis device according to claim 7, characterized in that, The device is deployed in a distributed cluster environment, supports horizontal scaling and load balancing, and has high availability and high concurrency processing capabilities.

Images