Technology System

◇  Provides HBase and HDFS engines for large scale raster data.

◇  Provides Elasticsearch engines for streaming data.

◇  Supports customized extensible geospatial big data storage engine.

◇  Kernel level extended Spark geospatial data model.

◇  Supports 6 categories covering 200 kinds of geospatial big data analysis operator.

◇  Supports building entities and relationships based on structured data.

◇  Supports associating graph entities with spatial attributes to map window for browsing.

◇  Supports graph query and analysis.

◇  Supports handling data lineage relationships for automated processes, writing them into graph databases for traceability and tracking.

◇  Integrated open source map development library (OpenLayers, Leaflet, MapboxGL JS) and visualization library (ECharts, MapV, DECK.GL).

◇  Provides 2D / 3D and dynamic / static visualization effects.

◇  Provides spatial big data visualization technologies such as scatter map, heat map, honeycomb map, grid map, trajectory map, O-D map, flow map, etc.

◇  Provides built-in spark clusters and external spark cluster.

◇  Supports Hadoop cluster YARN mode.

◇  Supports 4 Spark deployment modes: local (single-process), local (multi-process), client, and cluster.

◇ Supports processing of common satellite remote sensing data from GaoFen series, ZiYuan series, Beijing, GaoJing, Jilin, Sentinel, Landsat, IKONOS, WorldView, SPOT, and others.

◇ Supports managing intermediate and result data throughout the entire process in a mosaic dataset.

◇ Supports automatic generation of tie points and ground control points, enhancing geometric correction accuracy with AI semantics, and includes built-in AI heterogeneous image matching models to improve image matching accuracy.

◇ Supports block adjustment, offering linear, second-order polynomial, and third-order polynomial RPC correction models.

◇ Supports image registration, allowing for the registration of multispectral images based on geometrically corrected panchromatic images.

◇ Supports orthorectification, correcting the position and geometric distortions of original images based on adjusted RPC parameters.

◇ Supports image fusion algorithms such as Gram-Schmit and Pansharp, enabling users to control the sharpening effects of image fusion.

◇ Supports color output, synchronously adjusting image bit-depth and band sequence, enhancing output quality in vegetation areas.

◇ Supports image color balancing, using a reference image template to harmonize the colors of other images.

◇ Supports seamline generation, automatically avoiding linear features such as buildings, roads, or rivers.

◇ Supports output of result data in whole, standard subset, and polygonal clipping formats.

◇ Supports cloud removal, automatically replacing cloudy areas in image outputs.

◇ Supports haze removal, automatically enhancing contrast by removing thin haze and haze from images.

◇ Supports point cloud generation by performing dense matching on stereo images, producing dense point cloud data. Includes built-in AI dense matching models to effectively reduce mismatches.

◇ Supports Digital Surface Model (DSM) generation based on stereo images.

◇ Supports automatic filtering of DSM to generate Digital Elevation Model (DEM) products, utilizing AI filtering models that are particularly suitable for DSM filtering in urban areas.

◇ Supports DSM merging, combining multiple tiles of data into a single DSM covering the entire area.

◇ Supports planar and seam accuracy checks, displaying inspection points through thematic maps for visual assessment of result accuracy.

◇ Supports regional accuracy checks, facilitating detailed zone inspection of entire images.

◇ Supports deformation detection, automatically extracting areas with terrain, road, and building deformations.

◇ Supports generic radiation calibration models.

◇ Supports atmospheric correction for satellite sensor data from GF-1, GF-2, GF-6, GF-7, ZY3-02, etc.

◇ Supports image interpretation tasks including scene classification, change detection, object detection, binary classification, object extraction, and land cover classification across various scenarios.

◇ Provides ready-to-use pre-trained models for typical scenes like forest, farmland, urban water bodies, and urban buildings.

◇ Supports Large Imagery Model (LIM) for high-precision land cover classification in remote sensing interpretation.

◇ Supports G-SAM for fine-grained image segmentation by incorporating semantic information through spatial cues.

AI workflow tool

◇  Supports batch generation of training data during the data preparation stage.

◇  Supports training of multi-band remote sensing image interpretation models, along with capabilities for transfer learning and continuation training during the model construction stage.

◇  Supports various inference scenarios such as batch processing and range-based processing. Also includes capabilities for model eva1uation and conversion during the model application stage.

AI remote sensing image processing

◇  Built-in AI heterogeneous image matching models enhance the accuracy of matching different source images.

◇  Built-in AI pre-trained models for remote sensing interpretation utilize semantic information to remove non-ground points distributed in cloud-covered or built-up areas, enhancing geometric processing accuracy.

◇  Built-in AI dense matching models effectively reduce mismatches, improving the quality of Digital Surface Model (DSM) outputs.

◇  Built-in AI filtering models adapt to complex terrain conditions, particularly suitable for DSM filtering in urban areas.

AI Remote Sensing Image Interpretation

◇  Supports binary classification, land cover classification, general change detection, object detection, object extraction, scene classification, and other multi-scene image interpretation.

◇  Provides pre-trained models for various land covers such as forest land, farmland, urban water bodies, and urban buildings.

◇  New remote sensing interpretation pre-trained large model (LIM) capable of high-precision land cover classification.

◇  Supports G-SAM, enabling fine-grained image segmentation and the incorporation of semantic information through geographical prompts.

AI 3D Data Processing and Analysis

◇  Provides functionality for precise instance segmentation of point clouds, enabling accurate identification of objects such as power lines and towers.

◇  New capabilities for automatically extracting building outlines and identifying roof details, supporting the bulk construction of complex building entities.

AI Spatial Analysis

◇  Supports clustering analysis, including spatial hotspot analysis and spatial density clustering.

◇  Supports classification analysis capabilities, including map matching, address feature recognition, and forest-based classification.

◇  Supports regression analysis features such as geographically weighted regression, linear regression, and logistic regression.

AI Image/Video Analysis

◇  Supports image classification and object detection.

◇  Supports video Analysis, including capabilities for object detection, object tracking, object counting, speed analysis, geofencing, fight detection, license plate recognition, and supports custom sample training for video object detection models.

AI Image Generation

◇  New image generation feature that supports custom prompt words. It combines multiple reference images with various attribute constraints to generate intended images. These constraints include: sketch, outline, depth, line segment, semantic information, doodling, color, and lighting effects.

◇  New capability for local redrawing with support for custom prompts. This feature enables users to combine smudging, clicking, or text prompts to segment target areas for redrawing.

◇  New technical capability for super-resolution reconstruction, enabling customizable high-definition image enlargement using AI algorithms to enrich fine details in the image.

◇  New technical capability for fine-tuning image generation styles based on a small number of target style images. This feature allows for customizing generated image styles through refined model training.

AI Knowledge Generation

◇  New intelligent Q&A function for supporting batch upload of typical text materials such as doc, excel, pdf, etc., to build industry knowledge bases or provide fixed URLs. Based on a large NLP model, it enables the creation of custom scenario Q&A and supports answer tracing.

◇  Knowledge Summarization Functionality: Adds knowledge summarization function for uploading lengthy text materials. Based on a large NLP model, it provides functionalities such as summary generation, meeting minutes generation, topic analysis, long-text Q&A, and text profiling.

Geospatial AI Agent: SuperMap AgentX

◇  New technical capability for a geographic spatial intelligence agent. This feature enables users to submit spatial data processing requests through dialogue. It leverages advanced natural language processing models and AI agent technology to autonomously analyze, plan, and decompose tasks. Additionally, it interfaces with specialized spatial intelligence software tools to achieve specified objectives.

SuperMap iServer & iDesktopX

◇  Supports functionalities including AI remote sensing image processing, remote sensing image interpretation, spatial analysis, automated AI-driven 3D model construction, image analysis, and video analysis.

SuperMap ImageX Pro/Enterprise

◇  Supports capabilities such as AI remote sensing image processing and interpretation

SuperMap iPortal

◇  Supports interactive dialogue-based data resource retrieva1, analysis, visualization, portal management, and third-party functionality expansion through GI Agent.

SuperMap iMobile

◇  Supports AI capabilities for object recognition, detection, tracking, and counting.

◇  Supports AI mapping capabilities for indoor environments.

◇  Industry-specific applications in natural resource information can leverage various AI capabilities provided by SuperMap AIF. These include custom development of products such as intelligent Q&A systems for natural resources based on AI knowledge generation and spatial planning AI image generation applications.

◇  Supports distributed geospatial file system, including HDFS, DSF, etc.

◇  Supports distributed SQL geospatial database.

◇  Supports distributed NoSQL geospatial database, including MongoDB, Elasticsearch, HBase, etc.

◇  Supports federated storage of blockchain geospatial data Fabric and IPFS.

◇  Supports blockchain certificate service management capabilities.

◇  Provides kernel level extended Spark geospatial data model.

◇  Supports more than 200 kinds of distributed processing automation modeling tools.

◇  Provides more than 70 kinds of distributed geospatial analysis operators.

◇  Provides the high-performance distributed dynamic rendering capability.

◇  Realizes the full microservice of map, 3D, big data,AI and stream data functions.

◇  Supports docker deployment and provides rolling upgrade, elastic scaling and error recovery of GIS node.

◇  Provides general automatic layout based on Kubernetes, which can realize the real time monitoring of all microservice resources.

◇  Integrated service grid technology supports gray publish, access control, service measurement, and service tracking.

◇  Integrates serverless technology, and supports running spatial analysis tasks, geographic processing automation tasks, distributed image cutting tasks and 2D/3D tile distribution tasksin a Serverless GIS function mode.

◇  Supports the use and conversion of cloud native GIS data formats and tile formats such as FlatGeobuf and PMTiles.

◇  Provides edge GIS products to build cloud, edge and terminal integrated application system.

◇  Supports pre-proxy, service aggregation, data distribution and analysis processing of edge GIS.

◇ CPUs: x86, ARM, MIPS, SW-64, etc.

◇ Operating systems: Linux, Windows, HarmonyOS, Android, iOS, etc.

◇ Supports database series: file type, relational type, NoSQL, etc.

◇ Cloud platform: public cloud, private cloud, etc.