​​​​​​​

مرکز پردازش سریع (HPC)

​​​​​​​The High-Performance Computing (HPC) Center at the Faculty of Surveying and Geospatial Engineering is introduced as a relatively capable hardware infrastructure, designed to provide specialized support for research projects, student theses, and the academic work of researchers in the field of Geomatics Engineering.

Given the data-intensive and computationally demanding nature of modern research in spatial sciences, this center leverages high-performance computing technologies to offer an ideal platform for processing large volumes of satellite imagery, point clouds, 3D modeling, and other heavy computational tasks.

Established in early 2023 with the acquisition of three high-specification computers, the center was initially created to assist students with processing data for their theses and course projects. Since then, the number of servers has grown to five. Access to these resources has significantly alleviated the hardware constraints students previously faced in their processing workflows. The establishment of this facility has notably reduced the need for students to rent external servers at high costs, addressing a long-standing challenge within the faculty.

🛰️ Importance of the HPC Center for Geomatics Students:

1- Processing Large-Scale Spatial Data:

    Simultaneous processing of thousands of satellite images (e.g., Sentinel, Landsat, and high-resolution imagery) for projects involving change detection, land cover classification, and feature extraction.

    Processing point clouds derived from LiDAR scanners and UAVs, with volumes ranging from several gigabytes to terabytes.

    Execution of photogrammetric algorithms for generating Digital Surface Models (DSM) and Digital Terrain Models (DTM).

2- Advanced GPU Computing:

    Utilization of Graphics Processing Units (GPUs) to accelerate deep learning workflows in remote sensing, including object classification, feature identification, and change detection.

    Fast execution of complex algorithms in software environments such as MATLAB, Python (with TensorFlow and PyTorch libraries), and ENVI.

3- Environmental Simulations:

    Running hydrological and climatic models, along with spatial analyses for projects related to natural resource management, climate change studies, and urban planning.

:Specialized Services for Geomatics Students and Researchers👨‍💻  

Dedicated User Accounts: Graduate students and researchers working on computationally intensive projects are eligible to receive sufficient storage space and computing resources, provided they submit a completed server request form approved by their thesis supervisor and the faculty’s research deputy. Allocation is carried out in order of request date, and access is granted upon meeting the necessary requirements.

Technical Support and Consultation: The center offers expert guidance on code optimization and migration to parallel processing environments, while ensuring stable connectivity and minimizing runtime interruptions.

Software Library: A comprehensive suite of specialized geomatics software, including ENVI, PCI Geomatics, ESRI ArcGIS Pro, and other industry-standard tools, is pre-installed and configured to support the computational needs of various research projects.

🎯 Examples of Practical Applications for Student Theses and Dissertations:

- Rapid time-series analysis of satellite imagery for drought monitoring and land use/land cover change detection over specified time periods
- Automated semantic segmentation of very high-resolution (VHR) images using deep learning techniques
- Investigating drought impacts on forests and distinguishing them from deforestation using satellite data and machine learning
- Bayesian approaches for integrating and predicting water storage changes based on satellite and model-based data
- Enhancing the spatial resolution of UAV-derived orthophotomosaics
- Fusion of aerial and terrestrial imagery for improved 3D city model reconstruction
- Monitoring, analysis, and prediction of land subsidence using deep learning based on the LiCSBAS technique
- Geometric assessment of linear features to improve accuracy components in volunteered geographic information (VGI)
- Estimation of soil organic carbon content using remote sensing imagery
- Intelligent mitigation of urban heat islands using machine learning-based climate models
- Modeling and forecasting urban land use changes based on urban sprawl indices
- 3D city modeling through parallel processing of UAV imagery and point clouds
- Application of deep learning algorithms for automated detection of buildings, roads, and vegetation in VHR satellite images
- And many other research topics suitable for graduate theses and dissertations