The present invention will be further described in detail below with reference to the accompanying drawings and specific examples.
Such asfigure 2 As shown, a remote sensing image processing system based on a plug-in scalable architecture includes the following program modules running on the server:
Remote Sensing Plug-in Load Module: For the system's kernel module, based on the software profile search specified directory, according to the system-defined plugin rules, search for the DLL file in the directory, search for the eligible plugin file, start the priority, dependence on the plugin The relationship is analyzed, building a plug-in relying on the tree, and starts various plugins according to the dependencies of the plugin;
Interface Support Module: Provides a man-machine interface support service required for system operation, including system main program interface, buttons, map controls, floating window controls, remote sensing controls, and vector elements, providing services through API ways, providing a service for plugins The corresponding interface;
Communication Service Module: Provides a communication service between the platform and the plug-in, the communication service between the plug-in and the plug-in is provided in a message flow and data stream, and the message stream is a data set containing a single or multiple execution commands, and the data stream is a remote sensing image data. Vector data, data set to be processed by text parameters; the system will pass the message stream and the data stream, with high stability and scalability.
Remote Sensing Plug-in Life Cycle Management Module: Received the plugin information sent by the remote sensing plug-in load module, complete the management of the life cycle of the plugin, the management of the plugin status includes identification state, analyzing, start, activation, stop, unloading six types State, six states can be transformed from each other;
Log Management Module: Provide log management services, divide the log into general information, debug information, warning information, error message, and fatal error 5, providing services in the API form, plugin calls the corresponding service completed log information formatted record;
Remote Sensing Data Analytical Module: Complete the analysis of various types of remote sensing data, including raster image data and vector data, used to resolve various data formats, acquire metadata information of data, implement the data content read, and in the form of API Provides a service to the outside; the module can be plug-in expanded in various elements such as satellites, data formats, and data bits, with strong scalability.
Remote Sensing Algorithm Support Module: Provides management functions of remote sensing algorithms, including algorithm registration, algorithm uninstallation, algorithm monitoring, and algorithm maintenance, classification management in accordance with human machine interaction algorithm and automation processing;
Coordinates and Projection Conversion Modules: Complete different coordinate systems and conversion functions of different projection methods such that data of different coordinates and projection systems can be processed and displayed in the same coordinates and projection methods;
Image Pyramid Generating Module: To achieve a smooth image browsing effect, the image pyramid generating service is provided for wide remote sensing images, high-resolution remote sensing images, according to the size of the image, automatically determine the grade of the image pyramid, generate the corresponding pyramid, in the image When browsing, the corresponding pyramid data is automatically selected according to the browsing area of the image;
Image denoising module: Complete the noise removal of optical images, SAR images, enhance the application effect of the image, with Gaussian noise, pepper noise, striped noise processing ability;
Geometric calibration module: Complete an image correction based on RPC information, remember the image data according to the selection coordinate system and projection method, generate a secondary remote sensing image product;
Image Fusion Module: The input full-color image and multi-spectral image use the Pansharpen processing algorithm for fusion processing, generated remote sensing images, the spatial resolution is consistent with the full color image, the spectral resolution is consistent with the multi-spectral image;
Splicing Mosaic Module: Splicing the plurality of images adjacent or overlapping in a geographic location, first determine the position of the image coincidence according to the geographic location information of the image, and then extract the SIFT feature point, calculate multiple images. The same name feature point pair, based on this, the inlaid line is calculated, and the image of the double linear difference algorithm is calculated using the image of the coincidence region, and then the splicing area is subjected to a feathering process;
Image enhancement module: Improve image visualization effect by stretching grayscale, supported enhancement mode includes linear 2% stretching, gamma parameter stretching, histogram balance, histogram matching;
Target Detection Module: Fusion Image generated by the input full color and multi-spectral image or video fusion module, based on target detection of the target detection depth learning model, and outputs the position coordinates of the target external rectangle frame, support detection target The species include ships, aircraft;
Image Classification Module: Fusion images generated for the input full color and multi-spectral image or video fusion module, based on image classification based on image classification depth learning model that has completed training, and outputs image classification results; support classification target species include water body, Vegetation, soil, building, road.
The plug-ins will be described by XML document, describing the name, version, initialization status, start level, dependency information, activation strategy, and extension information of the plugin, the plug-in specification, plugin communication specification, plug-in management specification, plugin designation, plug-in management specification, plug-in description For four aspects of the norm, in which the plug-in interface includes three types of interface interfaces, service interfaces, management interfaces: interface interface provides interface elements generation and response function, each function plugin can add buttons and toolbars to the platform, service interface is used to provide The function of the communication service, meets the communication requirements between the platform and plugin, plug-ins, and the plug-in, and the management interface has openness and versatility, so that the application framework can use multiple development packages and support multiple data sources; plug-in communication specification uses messages The communication between the plug-in, separates the instructions and data, clearly divided into both messages and data, and implements multiple computing modules in parallel computing by message delivery mechanism; plug-in management specifications include Seven Basic Steps: Find the plugin file, plug-in preload, plugin validity check, identify search plugin dependent on the specified location and make preload, plugin file, generate plug-ins custom interface elements, use and setup Plugin.
Such asimage 3As shown, the life cycle of the plug-in includes the identification state, analyzes, start, the activation state, stop, the transfer of six states, the conversion rules of the six states are as follows: The plug-in framework first checks the plugin before installing the plugin, After the check is changed, the plug-in state is changed. After the plugin is installed, the plug-in framework parses the plugin resources and dependencies, the insert is successful, the plug-in state is changed, and the failure is still recognized; the plugin is resolved, after the plugin framework The plug-in can be activated, and the plug-in start operation is called by calling the plug-in activator and starts the plugin according to the activation policy. After the start is successful; the plugin becomes the active state; the plugin starts to the startup completion process, the plugin is in the state; the plugin stops by call The STOP method of the plug-in activator is stopped, the plug-in stops back to the resolver; during the plugin stops, the plugin is stopped in the stop; the plugin is stopped after the stop is completed; the unloading of the plugin is implemented by the Uninstall method, after the uninstallation is complete The plugin becomes a unload state.
Such asfigure 1 As shown, the module is built in:
a) Using a platform - plug-in architecture;
b) Functional system is divided into three levels: platform function, basic plug-in, handling plugin;
c) The platform can load plug-ins for a standard specification;
d) Full life cycle management of the platform for plugins;
E) The platform can perform the hot swap of the plugin.
In this system, the microennote platform performs plug-in management and communication management functions, with the plug-in management function completes the management of the full lifecycle of the infrastructure and the process plugin, etc., loading, run, unloading, etc. Communication between plugins and plugins; basic plug-in features include user interactive interface, remote sensing data resolution, coordinate projection conversion, log management 4 functions, with plug-in expansion ability; processing plugin includes image pyramid generation, image denoising, Geometric calibration, image fusion, splicing mosaic, image enhancement, target detection, image classification 8 processing functions, with plug-in expansion capabilities. The platform function can be called by the base plugin and processing plug-in, the base plugin function can be processed, and the processing plugin function can be called each other.
In summary, the present invention is aimed behind the development of the remote sensing information processing information system, and the system architecture model of "platform-plug-in" is used, and the system architecture model of "platform-plug-in" is used, and the processing application requirements for the management of remote sensing information is used. The status quo and development trend, using a series of advanced technologies such as platform / plug-in, constructing information system entities with universal technology architecture and good scalability, with flexible grouping, dynamic loading operation capacity, display remote sensing information sharing, resource co-utility Provide technical support and achievement foundation for the application of remote sensing information.