Next Article in Journal
Drivers of Farmers’ Willingness to Recycle Pesticide Packaging Waste: A Configurational Analysis
Previous Article in Journal
ESG Score and Firm Performance: A Comparative Analysis of Nordic and European Companies
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Sustainability
Volume 18
Issue 4
10.3390/su18041709
sustainability-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

A Co-Designed High-Performance Computing and Visualization Framework for Near-Real-Time Sustainable Land Governance Decisions

by
Pengfei Cong
1,2,†,
Mingxuan Yi
1,2,†,
Shibao Deng
3,*,
Qian Xiao
4,
Xinfeng Wang
1,
Wenmiao Zhao
1,
Chong Liu
1,
Yan Zhang
1 and
Jichao Gao
1
1
Langfang Integrated Natural Resources Survey Center, China Geological Survey, Langfang 065000, China
2
Innovation Base of Natural Resources Change Observation and Capital Monitoring in the Northern Haihe River Basin, China Society of Territorial Economists, Langfang 065000, China
3
Anhui Chemical Geological Exploration Institute, Anhui Geological and Mineral Exploration Bureau, Ma’anshan 243000, China
4
Beijing Shuhui Shikong Information Technology Co., Ltd., Beijing 100071, China
*
Author to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Sustainability 2026, 18(4), 1709; https://doi.org/10.3390/su18041709
Submission received: 2 January 2026 / Revised: 21 January 2026 / Accepted: 5 February 2026 / Published: 7 February 2026
Browse Figures
Versions Notes

Abstract

Sustainable land governance requires timely and accurate monitoring of land-use change to balance ecological, agricultural, and urban demands. Yet policymakers rarely receive actionable insights fast enough, because large-scale geospatial computation and rapid delivery remain disconnected. To close this gap, we introduce a Computational-Visualization Co-design (CVC) framework that welds a distributed high-performance computing engine to a real-time, preprocessing-free visualization system. Our approach represents a system-level innovation. It co-designs computational shards as visualization units, eliminating intermediate data reorganization. This co-design paradigm makes analytical results immediately visible. CVC processes a 20 TB imagery dataset and overlays millions of parcels 5–9 times faster than conventional engines. Map service publishing plummets from 168 h to just 7—a 24-fold speed-up—while client-side performance stays robust. The framework directly supports sustainable land management. It enables proactive monitoring, rapid impact assessment, and evidence-based policy formulation. Our work thus contributes to key Sustainable Development Goals related to land and cities. Validated with national survey data from China, the system merges analysis with instantaneous visual feedback, offering a practical route to sustainable land governance.

1. Introduction

Sustainable governance of land resources is fundamental to global food security, climate stability, biodiversity, and equitable urban growth [1,2]. Effective stewardship hinges on the capacity of governments and agencies to accurately detect and swiftly respond to changes in land use and cover [3,4]. This need for timely insight is amplified by global sustainability commitments, including the UN Sustainable Development Goals on sustainable cities and terrestrial ecosystems [5,6]. Transitioning from periodic, backward-looking reports to continuous, near-real-time monitoring is therefore not merely beneficial—it is a practical necessity. Such a shift enables proactive interventions, from curbing unlawful deforestation and improving farmland use to guiding responsible urban development [7,8]. Furthermore, even legally sanctioned changes in land use are often approved and implemented at decentralized local levels (e.g., county or city). The collective impact of these dispersed decisions may not be immediately apparent to provincial planners, creating a critical information gap. This lack of a unified, real-time overview impedes strategic coordination and sustainable governance at the regional scale.
A persistent technological bottleneck underlies this limitation. Current systems for monitoring land change often suffer from fragmented workflows [9,10], a problem rooted in the parallel development of two largely separate technology strands. One strand focuses on high-performance geocomputation, employing parallel frameworks [11], optimized algorithms [12], and distributed data management [13,14,15] to accelerate analytical processing. The other advances geospatial visualization through methods such as dynamic tiling [16], enhanced web services [17,18], and efficient raster pyramids [19,20]. This divergence points to a deeper methodological split. Researchers often treat computational scalability and interactive visualization as separate goals [21,22]. While each has its strengths, the two remain disconnected in practice. Consequently, even fast analytical results must go through outdated map-serving pipelines. These pipelines rely on slow, preprocessing-intensive tile generation [23,24]. This creates a critical delay, often lasting hours or days, between completing an analysis and delivering it in a visual, actionable form to decision-makers [25,26]. Such latency undermines the responsiveness required for sustainable governance, where policies must adapt to rapidly changing conditions [27,28]. Moreover, most existing solutions are also generic; they fail to address the specific demands of land administration, which include complex parcel geometries, strict topological rules, and legally mandated workflows [29,30,31].
To bridge this gap, we introduce and validate a Computational-Visualization Co-design (CVC) framework. This work moves beyond simply integrating two independent systems. Instead, we propose a new system design paradigm. In this paradigm, the basic units of computation, such as data shards, are simultaneously designed as the core indexing units for visualization. This architectural co-design eliminates the fundamental bottleneck of intermediate data reorganization, which plagues traditional “optimize-then-connect” approaches. The outcome is a seamless pipeline where analytical results become immediately visible for land-change verification.
The primary contribution of this work is therefore a system-level architecture founded on a co-design principle. This principle ensures that computational and visual processes are mutually reinforcing from the ground up. We materialize this through three key advances: an adaptive computing engine whose data shards are visualization-ready; a real-time visualization engine that operates without preprocessing bottlenecks; and a rigorous validation using national-scale data. Together, they demonstrate a 5–9× speedup in processing and a 24× faster time-to-visualization. This work provides more than faster components; it offers a proven blueprint for a new, responsive paradigm in sustainable land governance.

2. Materials and Methods

Addressing the disconnect between geocomputation and decision support requires a fundamental re-engineering of the land verification workflow. This study presents an integrated framework that co-designs high-performance distributed computing with a real-time visualization system. The following sections detail this architecture, beginning with an overview of the unified pipeline before delving into its core computational and visual engines.

2.1. System Overview and Integrated Workflow

Conventional verification systems often compartmentalize data processing, analysis, and visualization into sequential, batch-oriented steps, creating latency at each stage. Our framework was designed to collapse these stages into a cohesive, near-real-time pipeline. As depicted in Figure 1, the system architecture is built around a continuous flow from data ingestion to interactive decision support.
The process initiates with the automated ingestion and service publishing of incoming land survey data. Upon receipt, data are versioned and cataloged in a central repository. Critically, rather than relegating service creation to a later, manual step, the framework immediately publishes both raster imagery and vector parcels as standard web map services (WMS/WMTS). This “service-upon-ingestion” capability, driven by predefined metadata rules, ensures that the latest data are instantly discoverable and queryable, eliminating a traditional preprocessing bottleneck.
Once services are live, an automated task dispatch mechanism parcels the data and generates corresponding analysis jobs, distributing them across a distributed computing cluster. This automation ensures efficient resource allocation and sets the stage for large-scale processing.
The core analytical heavy lifting is performed by the distributed overlay processing engine. To manage datasets that would overwhelm single machines, the system employs an adaptive data sharding strategy, intelligently dividing large geographic layers into smaller, parallelizable units. A high-performance spatial cluster then processes these units concurrently, with all results being streamed back in real-time to a shared spatial database. This design not only accelerates computation but also makes results available the second they land for the next phase.
For the human-in-the-loop, a web-based interactive verification interface provides the front end. This interface automatically loads the processed verification targets from the shared database, presenting them clearly for expert review. To provide essential spatial context during this review, the system dynamically streams base map layers on demand. By leveraging dynamic tiling and sparse pyramid techniques (detailed in Section 2.3), the system delivers current imagery and vector backgrounds without pre-generation delays, allowing inspectors to work with a complete, up-to-date visual picture.
In essence, this architecture integrates data management, parallel computation, and instant visualization into a fluid workflow. It transforms a traditionally fragmented, slow process into a responsive system where insights can be computed and visually accessed on operational timescales.

2.2. Distributed Parallel Computing Engine

Conventional GIS architectures struggle with the computational scale required for overlay analysis of land parcels, a core task in change verification that often involves millions of complex features. To meet the demands of operational workflows, we developed a dedicated distributed computing engine. Its design integrates three key components: a distributed storage backend, an adaptive data sharding strategy, and an intelligent task scheduler. Together, these components shift overlay analysis from a sequential, memory-limited process into a parallelized, cluster-scale operation.
We used GeoPandas [32,33] for core spatial operations, taking advantage of its combined support for tabular data handling via Pandas and geometric operations via Shapely. This allowed us to perform the necessary range of tasks, from simple spatial queries to compute-intensive overlays. To support distributed storage and execution, we implemented Citus 13.0 [34], a PostgreSQL extension that partitions data horizontally across a cluster and provides the foundation for parallel processing. The choice of these tools was guided by several critical factors. GeoPandas delivers a mature and expressive framework for spatial operations in Python 3.8, effectively combining the geometric processing capabilities of Shapely with Pandas’ data analysis features. This integration makes it well-suited for developing and implementing complex spatial overlay workflows. We selected Citus as our distributed PostgreSQL extension due to its full support for SQL and GIS standards, along with its ability to transparently shard data across a cluster—ensuring strong consistency while building upon the reliable PostgreSQL foundation. For orchestrating distributed tasks, PowerJob provided a comprehensive open-source platform equipped with dynamic load balancing and robust fault tolerance. Its architecture demonstrated greater flexibility in managing our heterogeneous spatial task dependencies compared to conventional queue-based systems. Together, these components established an integrated and production-capable stack that optimizes development agility, operational stability, and horizontal scalability.

2.2.1. Adaptive Data Sharding

Traditional systems often process data as large, monolithic units. A common example is processing all parcels within an entire county at once. This approach can cause memory overflow and severe slowdowns during complex overlay analyses. Our approach proactively breaks these large spatial datasets into smaller, independent units called shards. During analysis, each node loads only the shards relevant to its assigned computation, significantly reducing memory and processing load per node. All results from parallel subtasks are written simultaneously to a shared database, which maintains data consistency while avoiding the limitations of single-node systems.
To further improve efficiency, the framework selects between two partitioning methods based on the analysis context. For tasks that follow administrative reporting lines, a hierarchical administrative division split is applied (Figure 2a). For analyses where boundaries are not meaningful, a regular grid-based partition is used (Figure 2b). This hybrid strategy balances managerial relevance with spatial efficiency.

2.2.2. Intelligent Distributed Scheduling

Effectively processing a large number of shards requires capable scheduling. We integrated an intelligent distributed scheduler that moves the workload from a single machine to a multi-node cluster. A dynamic load-balancing mechanism [35] lies at the center of this scheduler, continuously tracking CPU, memory, and I/O usage across all nodes in real time. It assigns tasks by considering both the computational difficulty of the job and the size of each shard, which helps avoid bottlenecks and idle resources, thus improving overall throughput.
The scheduler follows a centralized-decentralized model. A master node holds a global task queue and monitors worker nodes through persistent heartbeat signals. A dynamic scoring algorithm evaluates node readiness several times per second, allowing efficient management of billions of spatial shards. If a task fails, a layered retry process first attempts it again on the same node; if it continues to fail, the task is automatically moved to another available node. This ensures completion without manual intervention. Once processing is done, all worker nodes return their results to the master, where a streamlined aggregation algorithm merges them into a unified, decision-ready geospatial dataset, as shown in Figure 3.

2.3. Rapid Visualization Engine

A key challenge in operational land change monitoring is the delay between updating data and making it visually available. Traditional map-serving systems depend on pre-generated, static tile pyramids. Any change to the underlying data forces a complete regeneration of these tiles, a process that is time-consuming, storage-intensive, and prevents users from seeing updates right away [36]. To support instantaneous visualization, we developed a dynamic tile service based on a “service-upon-ingestion” model. This is paired with a sparse vector pyramid designed for the efficient display of very large parcel datasets.

2.3.1. Dynamic Raster Tiles for Instant Data Display

Our dynamic raster tile system reconstructs the serving pipeline by integrating on-demand tile generation, real-time spatial cropping, and parallelized rendering. As shown in Figure 4, the system works through three coordinated components. First, an automated service publisher creates standard WMS/WMTS interfaces directly during data ingestion, using predefined rules to make new imagery instantly discoverable. Second, a real-time cropping engine stores source raster data in efficiently partitioned blocks. When a user requests a map view, it dynamically retrieves and crops only the imagery within the requested extent and zoom level. Third, a high-performance rendering layer uses server-side parallel computing, load balancing, and on-the-fly compression to serve many concurrent requests without lag.
In practice, this means that new imagery can be served as soon as it arrives. As users pan or zoom, tiles are computed in real time for their viewport and streamed directly. This removes the need for large stores of pre-computed tiles, cuts storage requirements, and—most importantly—eliminates the re-tiling step from the update cycle. Land inspectors thus gain immediate visual access to the most current landscape, enabling decisions based on timely and accurate information.

2.3.2. Fast Vector Data Visualization Based on Sparse Vector Pyramids

Visualizing the vast number of parcels in national surveys requires an approach that balances detail with performance [37]. Following the same design philosophy as our raster system, we implemented sparse vector pyramid technology to enable instant display of hundreds of millions of parcels (Figure 5). Conventional vector tiling pre-generates uniform tiles for an entire region. This creates substantial data redundancy in areas with sparse parcel distribution, forcing clients to load unnecessary data and degrading overall performance [38].
Our sparse vector pyramid overcomes this by intelligently organizing data based on spatial density. During pyramid construction, a spatial indexing scheme first analyzes the distribution pattern of parcels. It then applies an adaptive tiling strategy: in high-density urban areas, it creates smaller, detailed tiles to preserve resolution, whereas in sparsely parceled rural or mountainous regions, it aggregates data into larger tiles to minimize the total tile count. This density-differentiated approach typically reduces the required tile generation and subsequent client-side processing load by approximately 50% compared to conventional uniform tiling.
The construction of the sparse pyramid is highly optimized, leveraging fully parallel processing [39] to build a hierarchical index for 100 million parcels in approximately ten minutes. For rendering, a client-side visual occlusion culling algorithm enhanced with Z-Order sorting [40] efficiently identifies and removes parcels outside the viewport or occluded by others. This ensures that only visibly relevant data is processed for drawing. When users zoom or pan, the system dynamically fetches only the necessary, non-redundant tiles from the sparse pyramid, enabling sub-second rendering and a smooth interactive experience even with continent-scale datasets.

3. Experimentation and Result Analysis

The following sections present the quantitative outcomes of evaluating the proposed integrated framework. We first detail the experimental setups designed to stress-test the system under realistic, large-scale land survey scenarios. We then report the benchmark results for the two core components: the distributed parallel computing engine for vector overlay analysis and the rapid visualization engines for raster and vector data publishing.

3.1. Distributed Parallel Computing Performance

3.1.1. Experimental Design and Data

To rigorously evaluate computational scalability, we designed a benchmark based on the most demanding operation in land change verification: the overlay analysis of complex cadastral parcels. Test datasets were extracted and synthesized from operational, multi-source county-level survey data. These datasets were structured in a graded series based on parcel count, ranging from 100,000 to 2.5 million features, as detailed in Table 1. A key characteristic of these datasets is their geometric complexity, featuring irregular shapes, multi-part polygons, and parcels with extremely high vertex counts (>10,000 nodes), accurately reflecting the computational challenges encountered in practice.

3.1.2. Computational Environment

All experiments were conducted on a dedicated, homogeneous computing cluster to ensure consistent and replicable measurements. The cluster architecture comprised one master node for coordination and ten worker nodes for parallel processing. The specific hardware configuration for each node type is provided in Table 2. The software stack used Powerjob for distributed task scheduling and dynamic load balancing. All nodes were connected via a high-speed 10-Gigabit Ethernet network. This setup minimized communication overhead [41,42]. A distributed PostgreSQL 14.0 database (Citus) served as the shared storage backend for both input data and computed results [43].

3.1.3. Overlay Processing Benchmark Results

The primary performance metric was the total wall-clock time required to complete a standard overlay analysis operation. We compared our distributed framework against a baseline established using a conventional, single-threaded GeoPandas 1.1.1 implementation, ensuring identical input data, overlay logic, and output quality [44]. The comparative results are summarized in Figure 6.
The baseline method exhibited a non-linear increase in processing time, culminating in a failure to complete the 2.5 million-parcel task due to memory exhaustion. This failure underscores the fundamental limitation of monolithic processing for operational-scale data. In stark contrast, the distributed framework successfully processed all datasets. The performance improvement factor scaled with data volume, reaching a 9× speedup for the 1 million-parcel dataset… This scalability stems primarily from the adaptive sharding strategy, which effectively bounds memory usage per node, and the dynamic scheduler, which optimizes resource utilization across the cluster. Most notably, the framework processed a dataset of 2.5 million parcels in approximately 30 min—a task that proved intractable for the baseline method. This result demonstrates the framework’s capacity to convert computations once considered infeasible or highly impractical into routine, operational tasks.

3.2. Rapid Visualization Performance

To evaluate the application effectiveness of rapid raster and vector visualization technologies in land change survey verification, this study focuses on data publishing efficiency and data browsing performance as core metrics. We designed comparative experiments to systematically analyze performance differences between traditional manual tiling and preprocessing-free tileless approaches, demonstrating the optimization potential of tileless rapid visualization over conventional workflows.

3.2.1. Experimental Design for Visualization

The visualization subsystem was evaluated on two critical, user-facing metrics: (1) Data Publishing Latency: The time elapsed from the moment a finalized dataset is ingested until it becomes available as a standard web map service (WMS/WMTS); and (2) Client-side Browsing Responsiveness: The map tile loading performance experienced by end-users under simulated concurrent access (Table 3).
For raster visualization, the test corpus was a nationally collected set of 434 very-high-resolution (0.5 m) satellite imagery scenes, totalling 20 TB. For vector visualization, tests were conducted at three administrative levels—county, city, and province—with the provincial dataset containing approximately 17 million parcels, characterized by high fragmentation across diverse terrains (Table 4). We simulated a multi-user environment based on the expected operational scale of a provincial land-resource agency. Browsing tests employed a sustained load of 30 concurrent virtual users. This number was set to approximate the peak count of specialist staff, such as data analysts and regional managers, who would likely require simultaneous system access for verification and decision-making. The hardware configurations for the raster and vector visualization servers are specified in Table 5 and Table 6, respectively.

3.2.2. Data Publishing Latency Results

The shift from preprocessing to on-demand rendering dramatically reduced the time required to publish data as interactive map services. As summarized in Figure 7, the performance gains were substantial for both data types.
For the massive 20 TB raster dataset, the dynamic tiling system cut the time-to-service from 168 h to just 7 h, achieving a 24-fold acceleration. The efficiency gains for vector data were equally significant and demonstrated clear scalability. Publishing a provincial-level dataset of 17 million parcels took only 30 min with our sparse vector pyramid approach, compared to 220 min using conventional pre-tiling—a 7.3× improvement.
These results concretely demonstrate the effectiveness of the “service-upon-ingestion” model, effectively eliminating the traditional map-publishing bottleneck and enabling near-real-time data updates.

3.2.3. Browsing Performance Results

Critically, the dramatic acceleration in data publishing did not degrade the interactive experience for the end-user. Under the 30-user concurrent load, the client-side performance of the dynamic systems matched or exceeded that of traditional pre-tiled services.
For raster data, the average tile loading time was 155 milliseconds for the dynamic renderer, statistically indistinguishable from the 150 ms measured for pre-generated static tiles (Table 7). This confirms that on-the-fly raster cropping and encoding introduced no perceptible delay.
For vector data, the efficiency of the sparse pyramid organization translated into faster client-side performance. The average tile load time for the dynamic vector service was 300 ms, which is 40% faster than the 500 ms required for serving conventional pre-tiled vector data (Table 8). This demonstrates that the tileless approach not only removes the publishing bottleneck but can also enhance the responsiveness of the final map service.

4. Discussion

The experimental results demonstrate that the proposed integrated framework delivers order-of-magnitude improvements in both processing throughput and data accessibility for land change verification. This discussion interprets these findings, articulates how they address the critical gaps identified in current systems, and positions the contributions within the broader context of high-performance geospatial computing.

4.1. Synergy Between Computing and Visualization

The key outcome is the synergy from co-design, not just faster individual parts. The distributed computing engine tackles the problem of analytical scalability, transforming overlay analysis from a memory-bound, often intractable, single-node task into a predictable, cluster-scale operation [45]. Crucially, this high-throughput analytical capability would be of limited operational value if its outputs remained trapped in a database, awaiting the slow, batch-oriented process of traditional map tile generation [18]. This is precisely the workflow disconnect noted in prior critiques of operational GIS.
Our rapid visualization engines directly dismantle this dissemination bottleneck [46]. The 24-fold reduction in raster publishing latency and the 7.3× acceleration in vector service generation effectively implement a “service-upon-ingestion” model [47]. When combined, these two advances create a seamless, end-to-end pipeline [48]. A complex, province-scale overlay analysis can now be computed and made available for interactive visual verification within approximately one hour—a timeframe that aligns with daily operational decision-making cycles and marks a substantive shift from retrospective reporting toward near-real-time monitoring [49].

4.2. System-Level Innovation Beyond Isolated Optimizations

Prior research has delivered excellent improvements in isolated areas. Examples include faster parallel overlay algorithms [50,51] and more efficient tile-serving systems [52]. However, these advances often operate in silos. They follow a traditional model: first optimize a component, then connect it to the next. Our framework is fundamentally different. Its novelty lies in holistic, system-level co-design. Here, architectural choices for computation and visualization are made together. They are interdependent and mutually reinforcing from the start [53].
Our adaptive data sharding strategy best exemplifies this philosophy. Each spatial shard has a dual purpose. First, it is a unit for balanced parallel processing. Second, it serves as a ready-made spatial index for the visualization engine. This dual role is intentional. It eliminates the costly step of reorganizing data between computation and visualization. Systems that simply bolt together pre-optimized modules [54] cannot avoid this overhead. Our co-designed approach does.
This approach challenges a common belief. Many think a dynamic, preprocessing-free system must trade client-side speed for agility. Our results prove otherwise [55]. For example, our sparse vector pyramid uses intelligent, density-based indexing. This allows client-side performance to not just match, but exceed that of static pre-tiling. The key is reducing data redundancy and client processing load [56]. This explains the 40% faster vector tile loading we observed.
Therefore, this work transcends the incremental optimization of individual components. It presents a systemic architecture where synergistic design choices—such as the unified data sharding strategy—enable high-performance computation and instantaneous visualization to co-elevate each other [57]. The outcome is a framework that delivers both the immediacy of real-time updates and superior interactive responsiveness, effectively bridging the capability gap identified in current land change verification workflows.

4.3. From Benchmarking to Operational Practice

Beyond technical benchmarks, this study validates the framework’s feasibility within the specific constraints of land administration [58]. The use of complex, real-world parcel data with intricate geometries and rich attribute schemas confirms that the system handles the “messy” data realities of the domain [59], not just synthetic benchmarks. The dramatic reduction in manual preprocessing—from days of tile generation to near-instant service availability—translates directly into tangible operational benefits: reduced labor costs, faster response to incoming data (e.g., new satellite imagery or field surveys), and the empowerment of analysts to perform iterative, exploratory verification with immediate visual feedback [60]. This addresses the pressing need from land managers for tools that are not only powerful but also agile and integrated into daily workflows [61].
The leap in performance also creates direct pathways to more sustainable land governance [62]. Publishing map data in 7 h instead of 168 h, for instance, can mean detecting illegal deforestation or unauthorized construction days sooner. This enables faster intervention to protect ecosystems and comply with land plans. Moreover, the near-real-time analysis loop supports evidence-based and adaptive policy-making. Authorities can now adjust strategies quickly in response to land change signals [63,64]. This responsiveness is key for balancing urban growth, farmland preservation, and ecological conservation. This framework was developed to address the specific operational challenges faced by land and natural resources management agencies, with a focus on the provincial level in China. Validated using real-world agency data, it is currently in a pilot implementation phase across several provincial survey centers. It directly supports Sustainable Development Goals (SDGs) 11 (Sustainable Cities) and 15 (Life on Land) by enabling near-real-time monitoring, which facilitates proactive ecological protection, efficient urban planning, and timely enforcement of land-use regulations.

4.4. Limitations and Future Directions

While the framework demonstrates robust performance, its current implementation suggests several promising avenues for future work. First, the adaptive sharding, though context-aware, relies on heuristic rules. Machine learning-driven scheduling algorithms could potentially analyze data distribution and query patterns in real-time to further optimize shard boundaries and task placement, especially for highly heterogeneous datasets. Second, although the framework’s architecture is inherently scalable, its decentralized application across multiple administrative levels, spanning national, provincial, and local jurisdictions, introduces significant challenges for data consistency and synchronized update propagation. Third, while the benchmark results demonstrate dramatic speedups, they are based on specific datasets and a single test run per configuration. Although we have observed consistent performance trends across other internal datasets, a more rigorous validation involving multiple, independent runs with statistical analysis would strengthen the generalizability of the performance claims. Key factors that can influence the measured processing times, especially for raster data, include the internal tiling structure of source imagery, network I/O performance during data shard distribution, and the level of concurrent cluster load. Consequently, resolving the ensuing federated governance and technical integration issues is pivotal to enabling true cross-level collaborative verification. Integrating lightweight blockchain or distributed ledger mechanisms could provide an auditable trail of data provenance and change history, enhancing trust and transparency in collaborative verification processes [65]. Finally, extending the dynamic visualization paradigm to encompass real-time 3D and augmented reality (AR) interfaces could further bridge the gap between digital analysis and physical field inspection, offering next-generation tools for land monitoring.

5. Conclusions

In summary, this study proposes and rigorously validates a co-designed high-performance computing and visualization framework that effectively addresses the persistent issue of workflow fragmentation in operational land-change monitoring. By architecturally integrating distributed computation with dynamic, on-demand visualization, we establish a new paradigm in which analytical outcomes become directly and immediately visible. This transforms land monitoring from a traditionally slow, batch-processed task into an interactive, near-real-time operation. We achieved speedups of 5× to 24× across the entire pipeline. This was done without compromising usability. This outcome clearly underscores the effectiveness of our synergistic co-design approach. Beyond land change verification, this work offers a practical and scalable blueprint applicable to other domains that demand rapid analysis and dissemination of large-scale, dynamic geospatial data, such as smart-city digital twins and emergency response systems.

Author Contributions

Conceptualization, P.C. and M.Y.; methodology, P.C., S.D. and Q.X.; software, M.Y. and W.Z.; validation, P.C., X.W. and Y.Z.; formal analysis, P.C., M.Y. and C.L.; investigation, P.C., X.W., W.Z. and J.G.; resources, S.D. and Q.X.; data curation, M.Y., C.L. and P.C.; writing—original draft preparation, P.C. and M.Y.; writing—review and editing, S.D., Q.X. and Y.Z.; visualization, W.Z. and J.G.; supervision, S.D.; project administration, P.C., S.D. and Q.X.; funding acquisition, S.D. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Department of Natural Resources of Anhui Province under the Anhui Provincial Natural Resources Science and Technology Project (Grant No. 2025-K-1). The APC was funded by the same grant.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Conflicts of Interest

Author Qian Xiao was employed by the company Beijing Shuhui Shikong Information Technology Co., Ltd. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

References

  1. Verburg, P.H.; van de Steeg, J.; Veldkamp, A.; Willemen, L. From land cover change to land function dynamics: A major challenge to improve land characterization. J. Environ. Manag. 2009, 90, 1327–1335. [Google Scholar] [CrossRef]
  2. Karthikeyan, L.; Chawla, I.; Mishra, A.K. A review of remote sensing applications in agriculture for food security: Crop growth and yield, irrigation, and crop losses. J. Hydrol. 2020, 586, 124905. [Google Scholar] [CrossRef]
  3. Gaur, S.; Singh, R. A Comprehensive Review on Land Use/Land Cover (LULC) Change Modeling for Urban Development: Current Status and Future Prospects. Sustainability 2023, 15, 903. [Google Scholar] [CrossRef]
  4. Gu, Z.J.; Zeng, M.M. The Use of Artificial Intelligence and Satellite Remote Sensing in Land Cover Change Detection: Review and Perspectives. Sustainability 2024, 16, 274. [Google Scholar] [CrossRef]
  5. Jin, G.; Deng, X.Z.; Chu, X.; Li, Z.H.; Wang, Y. Optimization of land-use management for ecosystem service improvement: A review. Phys. Chem. Earth 2017, 101, 70–77. [Google Scholar] [CrossRef]
  6. Yang, Q.; Liu, G.Y.; Casazza, M.; Dumontet, S.; Yang, Z.F. Ecosystem restoration programs challenges under climate and land use change. Sci. Total Environ. 2022, 807, 150527. [Google Scholar] [CrossRef] [PubMed]
  7. Xia, J.; Zhao, Z.Y.; Chen, L.Q.; Sun, Y.Z. How urban renewal affects the sustainable development of public spaces: Trends, challenges, and opportunities. Front. Environ. Sci. 2024, 12, 1482169. [Google Scholar] [CrossRef]
  8. Zhai, J.H.; Pu, L.J. Urban-Rural Differences in Cropland Loss and Fragmentation Caused by Construction Land Expansion in Developed Coastal Regions: Evidence from Jiangsu Province, China. Remote Sens. 2025, 17, 2470. [Google Scholar] [CrossRef]
  9. Yang, J.; Huang, X. The 30 m annual land cover dataset and its dynamics in China from 1990 to 2019. Earth Syst. Sci. Data 2021, 13, 3907–3925. [Google Scholar] [CrossRef]
  10. Alshayeb, M.J. Evaluating Ecosystem Service Trade-Offs and Recovery Dynamics in Response to Urban Expansion: Implications for Sustainable Management Strategies. Sustainability 2025, 17, 2194. [Google Scholar] [CrossRef]
  11. Magoulès, F.; Ahamed, A.K.C.; Putanowicz, R. Optimized Schwarz method without overlap for the gravitational potential equation on cluster of graphics processing unit. Int. J. Comput. Math. 2016, 93, 955–980. [Google Scholar] [CrossRef]
  12. Chang, C.C.; Lin, C.Y.; Wang, Y.Z. New image steganographic methods using run-length approach. Inf. Sci. 2006, 176, 3393–3408. [Google Scholar] [CrossRef]
  13. Almeida, P.S.; Shoker, A.; Baquero, C. Delta State replicated data types. J. Parallel Distrib. Comput. 2018, 111, 162–173. [Google Scholar] [CrossRef]
  14. Harder, D.; Shoushtari, H.; Sternberg, H. Real-Time Map Matching with a Backtracking Particle Filter Using Geospatial Analysis. Sensors 2022, 22, 3289. [Google Scholar] [CrossRef]
  15. Chen, Y.; Kar, S.; Moura, J.M.E. Resilient Distributed Parameter Estimation with Heterogeneous Data. IEEE Trans. Signal Process. 2019, 67, 4918–4933. [Google Scholar] [CrossRef]
  16. Hu, J.Y.; Luo, H.Z.; Jiang, H.; Xiao, G.Q.; Li, K.L. FastLoad: Speeding Up Data Loading of Both Sparse Matrix and Vector for SpMV on GPUs. IEEE Trans. Parallel Distrib. Syst. 2024, 35, 2423–2434. [Google Scholar] [CrossRef]
  17. Capolupo, A.; Monterisi, C.; Saponieri, A.; Addona, F.; Damiani, L.; Archetti, R.; Tarantino, E. An Interactive WebGIS Framework for Coastal Erosion Risk Management. J. Mar. Sci. Eng. 2021, 9, 567. [Google Scholar] [CrossRef]
  18. Netek, R.; Masopust, J.; Pavlicek, F.; Pechanec, V. Performance Testing on Vector vs. Raster Map Tiles-Comparative Study on Load Metrics. ISPRS Int. J. Geo-Inf. 2020, 9, 101. [Google Scholar] [CrossRef]
  19. Li, G.W.; Li, J.Z. An adaptive-size vector tile pyramid construction method considering spatial data distribution density characteristics. Comput. Geosci. 2024, 184, 105537. [Google Scholar] [CrossRef]
  20. Xie, S.; Liu, L.Y.; Zhang, X.; Yang, J.N.; Chen, X.D.; Gao, Y. Automatic Land-Cover Mapping using Landsat Time-Series Data based on Google Earth Engine. Remote Sens. 2019, 11, 3023. [Google Scholar] [CrossRef]
  21. Kerkad, A.; Gouri, R. Meta-visualization framework for spatiotemporal analytics: From data generation to advanced visualization on maps. Inf. Vis. 2025, 25, 41–65. [Google Scholar] [CrossRef]
  22. Wu, J.Y.; Gan, W.S.; Chao, H.C.; Yu, P.S. Geospatial Big Data: Survey and Challenges. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2024, 17, 17007–17020. [Google Scholar] [CrossRef]
  23. Huang, H.B.; Chen, Y.L.; Clinton, N.; Wang, J.; Wang, X.Y.; Liu, C.X.; Gong, P.; Yang, J.; Bai, Y.Q.; Zheng, Y.M.; et al. Mapping major land cover dynamics in Beijing using all Landsat images in Google Earth Engine. Remote Sens. Environ. 2017, 202, 166–176. [Google Scholar] [CrossRef]
  24. Li, Y. Monitoring and Mathematical Model Analysis of Dynamic Changes in Land Resources Based on SAR Sensor Image. J. Sens. 2021, 2021, 1661825. [Google Scholar] [CrossRef]
  25. Zhang, G.M.; Xu, J. Multi-GPU-Parallel and Tile-Based Kernel Density Estimation for Large-Scale Spatial Point Pattern Analysis. Isprs Int. J. Geo-Inf. 2023, 12, 31. [Google Scholar] [CrossRef]
  26. Jurik, V.; Herman, L.; Snopková, D.; Galang, A.J.; Stachon, Z.; Chmelík, J.; Kubícek, P.; Sasinka, C. The 3D hype: Evaluating the potential of real 3D visualization in geo-related applications. PLoS ONE 2020, 15, e0233353. [Google Scholar] [CrossRef]
  27. Zhao, J.B.; Liu, Y.X.; Yang, J.T. 3D matching positioning method for landslide using infrasound signal received by triangular pyramid vector array, based on ray theory. Bull. Eng. Geol. Environ. 2021, 80, 889–904. [Google Scholar] [CrossRef]
  28. Xing, H.Q.; Chen, J.; Wu, H.; Hou, D.Y. A Web Service-Oriented Geoprocessing System for Supporting Intelligent Land Cover Change Detection. ISPRS Int. J. Geo-Inf. 2019, 8, 50. [Google Scholar] [CrossRef]
  29. Tewkesbury, A.P.; Comber, A.J.; Tate, N.J.; Lamb, A.; Fisher, P.F. A critical synthesis of remotely sensed optical image change detection techniques. Remote Sens. Environ. 2015, 160, 1–14. [Google Scholar] [CrossRef]
  30. Ma, R.; Sun, E.D.; Zou, J. A spectral method for assessing and combining multiple data visualizations. Nat. Commun. 2023, 14, 780. [Google Scholar] [CrossRef] [PubMed]
  31. Babalola, K.H.; Hull, S.; Whittal, J. Alignment of the legal and institutional framework of Ekiti state land administration against the fit-for-purpose land administration framework principles. Surv. Rev. 2025, 57, 356–373. [Google Scholar] [CrossRef]
  32. Luo, Y.W.; Wang, X.L.; Xu, Z.Q. An agent approach to spatial information grid architecture design. Comput. Inform. 2005, 24, 201–222. [Google Scholar]
  33. Yan, X.J.; Liu, S.B.; Liu, W.; Dai, Q. An improved coverage-oriented retrieval algorithm for large-area remote sensing data. Int. J. Digit. Earth 2022, 15, 606–625. [Google Scholar] [CrossRef]
  34. Zheng, K.; Gu, D.P.; Fang, F.L.; Zhang, M.; Zheng, K.; Li, Q. Data storage optimization strategy in distributed column-oriented database by considering spatial adjacency. Clust. Comput. J. Netw. Softw. Tools Appl. 2017, 20, 2833–2844. [Google Scholar] [CrossRef]
  35. Dhakal, S.; Hayat, M.M.; Pezoa, J.E.; Yang, C.D.; Bader, D.A. Dynamic load balancing in distributed systems in the presence of delays: A regeneration-theory approach. IEEE Trans. Parallel Distrib. Syst. 2007, 18, 485–497. [Google Scholar] [CrossRef]
  36. Li, C.; Lu, H.M.; Xiang, Y.; Liu, Z.Q.; Yang, W.L.; Liu, R.L. Bringing Geospatial Data Closer to Mobile Users: A Caching Approach Based on Vector Tiles for Wireless Multihop Scenarios. Mob. Inf. Syst. 2018, 2018, 5186495. [Google Scholar] [CrossRef]
  37. Teijeiro, D.; Amor, M.; Doallo, R.; Corbelle, E.; Porta, J.; Parapar, J. Land consolidation through parcel exchange among landowners using a distributed Spark-based genetic algorithm. J. Supercomput. 2022, 78, 19522–19544. [Google Scholar] [CrossRef]
  38. Wang, S.N.; Liu, Y.; Li, L.L. Sparse Weighting for Pyramid Pooling-Based SAR Image Target Recognition. Appl. Sci. 2022, 12, 3588. [Google Scholar] [CrossRef]
  39. Han, H.; Han, Q.Q.; Li, X.J.; Gu, J.Y. Hierarchical spatial pyramid max pooling based on SIFT features and sparse coding for image classification. IET Comput. Vis. 2013, 7, 144–150. [Google Scholar] [CrossRef]
  40. Lee, K.C.K.; Lee, W.C.; Zheng, B.; Li, H.; Tian, Y. Z-SKY: An efficient skyline query processing framework based on Z-order. Vldb J. 2010, 19, 333–362. [Google Scholar] [CrossRef]
  41. Toptal, A.; Sabuncuoglu, I. Distributed scheduling: A review of concepts and applications. Int. J. Prod. Res. 2010, 48, 5235–5262. [Google Scholar] [CrossRef]
  42. Jian, H.Z.; Nee, A.Y.C.; Fuh, J.Y.H.; Zhang, Y.F. A modified genetic algorithm for distributed scheduling problems. J. Intell. Manuf. 2003, 14, 351–362. [Google Scholar] [CrossRef]
  43. Ilba, M. Parallel algorithm for improving the performance of spatial queries in SQL: The use cases of SQLite/SpatiaLite and PostgreSQL/PostGIS databases. Comput. Geosci. 2021, 155, 104840. [Google Scholar] [CrossRef]
  44. Zhou, Y.K.; Wang, S.H.; Guan, Y. An Efficient Parallel Algorithm for Polygons Overlay Analysis. Appl. Sci. 2019, 9, 4857. [Google Scholar] [CrossRef]
  45. Preston, M.; Clayton, P.; Wells, G. Dynamic run-time application development using CORBA objects and XML in the field of distributed GIS. Int. J. Geogr. Inf. Sci. 2003, 17, 321–341. [Google Scholar] [CrossRef]
  46. Sun, Y.Q.; Xu, C.; Wu, J.X.; Yuan, H.; Shi, H.B.; Zhan, X.M.; Zhang, L. Development of a multi-view and geo-event-driven real-time collaborative GIS. Int. J. Digit. Earth 2022, 15, 134–147. [Google Scholar] [CrossRef]
  47. Chen, Z.Q.; Chen, N.C. A Real-Time and Open Geographic Information System and Its Application for Smart Rivers: A Case Study of the Yangtze River. ISPRS Int. J. Geo-Inf. 2019, 8, 114. [Google Scholar] [CrossRef]
  48. Villarroya, S.; Viqueira, J.R.R.; Cotos, J.M.; Taboada, J.A. Enabling Efficient Distributed Spatial Join on Large Scale Vector-Raster Data Lakes. IEEE Access 2022, 10, 29406–29418. [Google Scholar] [CrossRef]
  49. Li, S.N.; Dragicevic, S.; Castro, F.A.; Sester, M.; Winter, S.; Coltekin, A.; Pettit, C.; Jiang, B.; Haworth, J.; Stein, A.; et al. Geospatial big data handling theory and methods: A review and research challenges. ISPRS J. Photogramm. Remote Sens. 2016, 115, 119–133. [Google Scholar] [CrossRef]
  50. Dritsas, E.; Trigka, M. Remote Sensing and Geospatial Analysis in the Big Data Era: A Survey. Remote Sens. 2025, 17, 550. [Google Scholar] [CrossRef]
  51. Zhao, K.; Jin, B.X.; Fan, H.; Song, W.W.; Zhou, S.Y.; Jiang, Y.Y. High-Performance Overlay Analysis of Massive Geographic Polygons That Considers Shape Complexity in a Cloud Environment. ISPRS Int. J. Geo-Inf. 2019, 8, 290. [Google Scholar] [CrossRef]
  52. Kong, M.; Veras, R.; Stock, K.; Franchetti, F.; Pouchet, L.N.; Sadayappan, P. When Polyhedral Transformations Meet SIMD Code Generation. ACM SIGPLAN Not. 2013, 48, 127–138. [Google Scholar] [CrossRef]
  53. Huang, K.J.; Wang, C.L.; Wang, S.H.; Liu, R.Y.; Chen, G.X.; Li, X.L. An Efficient, Platform-Independent Map Rendering Framework for Mobile Augmented Reality. ISPRS Int. J. Geo-Inf. 2021, 10, 593. [Google Scholar] [CrossRef]
  54. Zhang, J.B.; Ye, Z.Z.; Zheng, K. A Parallel Computing Approach to Spatial Neighboring Analysis of Large Amounts of Terrain Data Using Spark. Sensors 2021, 21, 365. [Google Scholar] [CrossRef]
  55. Xu, Y.Q.; An, J.W.; Zhou, C.C.; Xu, H.T.; Han, Z. Dynamic Energy Management for UAV-Enabled VR Systems: A Tile-Based Collaboration Approach. IEEE Trans. Veh. Technol. 2024, 73, 17668–17683. [Google Scholar] [CrossRef]
  56. Liu, J.J.; Wu, Z.B.; Sun, L.; Wei, Z.H.; Xiao, L. Hyperspectral Image Classification Using Kernel Sparse Representation and Semilocal Spatial Graph Regularization. IEEE Geosci. Remote Sens. Lett. 2014, 11, 1320–1324. [Google Scholar] [CrossRef]
  57. Rössmann, H.; Peyker, J.; Völker, A.; Klink, A. Application of Change Detection Methods for the Update of GIS Data of Sealed Surfaces and Buildings. Photogramm. Fernerkund. Geoinf. 2013, 5, 447–458. [Google Scholar] [CrossRef]
  58. Stefanovic, M.; Przulj, D.; Ristic, S.; Stefanovic, D.; Nikolic, D. Smart Contract Application for Managing Land Administration System Transactions. IEEE Access 2022, 10, 39154–39176. [Google Scholar] [CrossRef]
  59. Ma, M.Y.; Wu, Y.; Chen, L.; Li, J.; Jing, N. Interactive and Online Buffer-Overlay Analytics of Large-Scale Spatial Data. ISPRS Int. J. Geo-Inf. 2019, 8, 21. [Google Scholar] [CrossRef]
  60. De Beer, D.J.; Joubert, J.W. Evolutionary optimisation of large-scale activity clustering with increased automation. Comput. Oper. Res. 2022, 146, 105925. [Google Scholar] [CrossRef]
  61. Krizanovic, J.; Roic, M. Modeling Land Administration Data Dissemination Processes: A Case Study in Croatia. ISPRS Int. J. Geo-Inf. 2023, 12, 20. [Google Scholar] [CrossRef]
  62. Pickens, A.H.; Hansen, M.C.; Song, Z.; Poulson, A.; Komarova, A.; Baggett, A.; Kerr, T.; Mikus, A.; Ortiz Dominguez, C.; Tyukavina, A.; et al. Rapid monitoring of global land change. Nat. Commun. 2025, 16, 8948. [Google Scholar] [CrossRef]
  63. Kauark-Fontes, B.; Ortiz-Guerrero, C.E.; Marchetti, L.; Hernández-Garcia, J.; Salbitano, F. Towards Adaptive Governance of Urban Nature-Based Solutions in Europe and Latin America—A Qualitative Exploratory Study. Sustainability 2023, 15, 4479. [Google Scholar] [CrossRef]
  64. Woiwode, C.; Ramachandran, A.; Philip, T.; Rishika, D.; Rajan, S.C. Identifying entry points for adaptive governance in peri-urban Chennai (India): A multi-dimensional, multi-level, and multi-scalar approach. Front. Sustain. Cities 2024, 6, 1368240. [Google Scholar] [CrossRef]
  65. Yadav, A.S.; Kushwaha, D.S. Digitization of Land Record Through Blockchain-based Consensus Algorithm. IETE Tech. Rev. 2022, 39, 799–816. [Google Scholar] [CrossRef]
Sustainability 18 01709 g001
Figure 1. Architecture and workflow of the integrated CVC framework.
Figure 1. Architecture and workflow of the integrated CVC framework.
Sustainability 18 01709 g001
Sustainability 18 01709 g002
Figure 2. Illustration of the two data sharding methods (using the example of Liaoning Province, People’s Republic of China): (a) Data sharding by county-level administrative division codes; (b) Data sharding by a standard grid.
Figure 2. Illustration of the two data sharding methods (using the example of Liaoning Province, People’s Republic of China): (a) Data sharding by county-level administrative division codes; (b) Data sharding by a standard grid.
Sustainability 18 01709 g002
Sustainability 18 01709 g003
Figure 3. Distributed scheduling workflow.
Figure 3. Distributed scheduling workflow.
Sustainability 18 01709 g003
Sustainability 18 01709 g004
Figure 4. Dynamic raster tiles technical roadmap.
Figure 4. Dynamic raster tiles technical roadmap.
Sustainability 18 01709 g004
Sustainability 18 01709 g005
Figure 5. Fast vector data visualization technical roadmap.
Figure 5. Fast vector data visualization technical roadmap.
Sustainability 18 01709 g005
Sustainability 18 01709 g006
Figure 6. Overlay processing performance comparison between the baseline (single-threaded GeoPandas) and the proposed CVC framework. The baseline method fails at 2.5 million parcels due to memory overflow. The CVC framework successfully processes all datasets, achieving up to a 9× speedup and completing the previously intractable 2.5-million-parcel task in approximately 30 min. Note the logarithmic scale on both axes.
Figure 6. Overlay processing performance comparison between the baseline (single-threaded GeoPandas) and the proposed CVC framework. The baseline method fails at 2.5 million parcels due to memory overflow. The CVC framework successfully processes all datasets, achieving up to a 9× speedup and completing the previously intractable 2.5-million-parcel task in approximately 30 min. Note the logarithmic scale on both axes.
Sustainability 18 01709 g006
Sustainability 18 01709 g007
Figure 7. Data publishing latency comparison between traditional manual pre-tiling and the proposed CVC dynamic service for raster (20 TB) and vector (provincial, 17 million parcels) datasets. Note the logarithmic scale on the y-axis, highlighting the 24× and 7.3× acceleration for raster and vector data, respectively.
Figure 7. Data publishing latency comparison between traditional manual pre-tiling and the proposed CVC dynamic service for raster (20 TB) and vector (provincial, 17 million parcels) datasets. Note the logarithmic scale on the y-axis, highlighting the 24× and 7.3× acceleration for raster and vector data, respectively.
Sustainability 18 01709 g007
Table 1. Data sample characteristics table.
Table 1. Data sample characteristics table.
No.Feature Volume (Number of Patches)Data Characteristics
1100,000Contains irregular shapes such as nested inner and outer rings, multi-part combinations, etc. Single patch has over 10,000 nodes, complex topological relationships, and attribute fields generated from multi-source data overlay processing exceed 50.
2500,000
31,000,000
42,500,000
Table 2. Test Environment Component Configuration Parameters.
Table 2. Test Environment Component Configuration Parameters.
No.ItemConfigurationQuantity
1Master NodeCPU: 32-core, RAM: 64 GB, Disk: 2 TB1
2Compute NodeCPU: 8-core, RAM: 8 GB, Disk: 500 GB10
Table 3. Parameters for raster data quick display experiment.
Table 3. Parameters for raster data quick display experiment.
No.Data VolumeSpatial Resolution
1434 scenes/20 TB0.5 m
Table 4. Characteristics of administrative division data by level.
Table 4. Characteristics of administrative division data by level.
No.Administrative LevelFeature Volume (Patch Count)Data Characteristics
1County100,000Full coverage, complex terrain, high patch fragmentation
2City450,000
3Province17,000,000
Table 5. Configuration parameters for raster data rapid visualization test environment.
Table 5. Configuration parameters for raster data rapid visualization test environment.
No.CategorySpecifications
1Hardware Environment2 Linux servers (32-core CPU, 64 GB RAM, 2 TB storage)
2Concurrent Load30-user simultaneous access
Table 6. Configuration parameters for vector data rapid visualization test environment.
Table 6. Configuration parameters for vector data rapid visualization test environment.
No.CategorySpecifications
1Hardware Environment1 Linux server (40-core CPU, 128 GB RAM, 2 TB storage)
2Concurrent Load30-user simultaneous access
Table 7. Raster data browsing efficiency comparison.
Table 7. Raster data browsing efficiency comparison.
No.Browsing MethodSingle Tile Loading Time
1Static Pre-generated Tiles150 ms
2Rapid Raster Visualization155 ms
Table 8. Vector data browsing efficiency comparison.
Table 8. Vector data browsing efficiency comparison.
No.Browsing MethodSingle Tile Loading Time
1Manual Tile Processing500 ms
2Preprocessing-free (Rapid Visualization)300 ms
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Cong, P.; Yi, M.; Deng, S.; Xiao, Q.; Wang, X.; Zhao, W.; Liu, C.; Zhang, Y.; Gao, J. A Co-Designed High-Performance Computing and Visualization Framework for Near-Real-Time Sustainable Land Governance Decisions. Sustainability 2026, 18, 1709. https://doi.org/10.3390/su18041709

AMA Style

Cong P, Yi M, Deng S, Xiao Q, Wang X, Zhao W, Liu C, Zhang Y, Gao J. A Co-Designed High-Performance Computing and Visualization Framework for Near-Real-Time Sustainable Land Governance Decisions. Sustainability. 2026; 18(4):1709. https://doi.org/10.3390/su18041709

Chicago/Turabian Style

Cong, Pengfei, Mingxuan Yi, Shibao Deng, Qian Xiao, Xinfeng Wang, Wenmiao Zhao, Chong Liu, Yan Zhang, and Jichao Gao. 2026. "A Co-Designed High-Performance Computing and Visualization Framework for Near-Real-Time Sustainable Land Governance Decisions" Sustainability 18, no. 4: 1709. https://doi.org/10.3390/su18041709

APA Style

Cong, P., Yi, M., Deng, S., Xiao, Q., Wang, X., Zhao, W., Liu, C., Zhang, Y., & Gao, J. (2026). A Co-Designed High-Performance Computing and Visualization Framework for Near-Real-Time Sustainable Land Governance Decisions. Sustainability, 18(4), 1709. https://doi.org/10.3390/su18041709

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop