Speaker: Vaibhav Mehrotra (Inst. 193-03 PRIP)
College: Manipal Institute of Technology, Manipal Academy of Higher Education, India
Course: Bachelor of Technology in Computer and Communication Engineering
Semester: 8th Semester
Internship
As a part of the course curriculum, students at MIT Manipal are supposed to complete an internship as their final semester project either in the corporate or research domain. I have chosen to write my thesis in the field of computer vision and image processing under the guidance of Walter G. Kropatsch and other members of the Pattern Recognition and Image Processing Group at TU Wien.
The duration of the internship will be about 5 months, from 2nd January 2019 to 27th May 2019.
Topic and Abstract
Real-time geosensing has various commercial and social-welfare applications ranging from monitoring aquaculture to flood or landslide detection. These efforts can be hampered by cloud coverage. Synthetic Aperture Radar (SAR) uses frequencies in the microwave range which effectively penetrate clouds and provide grayscale intensity image of the landscape.
The aim of the project is to develop a model that can successfully segment and provide semantic labels to the gray-scale SAR images. Multiple labels will be taken into consideration including but not limited to land, water, buildings, vegetation and forest cover so that surface knowledge about any landscape can be derived.