Mohammad Salmani
I am a PhD student in Earth Sciences at the University at Buffalo. I specialize in applying remote sensing and GIS techniques to ice sheet modeling. My research focuses on utilizing Sentinel-2 data classification to determine the maximum ice sheet extent during the Little Ice Age across the entirety of Greenland.
- msalmani@buffalo.edu
- salmanii.mohamad@gmail.com
- Hochesteter Hall 471, University at Buffalo North Campus
- LinkedIn.com/in/mohammad-salmani-ut
Research Interests
- Remote Sensing of Ice Sheets
- Remote Sensing of ocean and monitoring of sea level rise
- Geographic Information Systems and Spatial Analysis
- Deep learning and Machine learning
About Myself
In 2019, I participated in the highly competitive Iranian University Entrance Exam and achieved a remarkable result, ranking among the top 2% of test-takers. Motivated by this success, I decided to pursue a degree in Surveying and Geospatial Engineering at the prestigious University of Tehran, marking the start of my journey in this field.
After completing my undergraduate studies, I was fortunate to secure a research assistant position at the University at Buffalo in 2024. This opportunity has allowed me to commence my doctoral degree program, which I am delighted to be undertaking under the supervision of Professor Beata Csatho. My current research focuses on identifying the trimline of Greenland from the Little Ice Age.
Programming Skills
Web Skills
Software Skills
My Research Projects
- Utilized the UNet++ architecture in deep learning to develop a change detection model for very high-resolution (VHR) images. The model effectively identifies changes in the given images.
- Employed the powerful AlexNet architecture to perform land use classification. Through this project, different land types were accurately classified, allowing for effective environmental analysis.
- Implemented various filters on a Landsat 8 image utilizing MATLAB, both in the spatial domain.
- Developed a Python code utilizing Affine, Conformal, Projective, and Polynomial transforms to accurately convert the coordinates of a selected point from a Landsat 8 image to its corresponding ground coordinates. This code enables precise geo-referencing and aligning spatial data with real-world locations.
- Designed and simulated a database inspired by the LinkedIn application using Oracle SQL Developer. Additionally, created a user-friendly Windows application utilizing C# to simplify the process of creating SQL queries. This enables more efficient querying and management of the database, enhancing the overall usability of the LinkedIn-like application.
- Enhanced the quality of a common object by co-registering two point clouds. By aligning the point clouds, the accuracy and resolution of the object were improved, facilitating detailed analysis, visualization, and potential applications in fields such as 3D modeling and object recognition.