Remote Sensing and Image Visualization Using Deep Learning( AERIAL IMAGERY SEMANTIC SEGMENTATION)


  • Machine Learning

  • Image Processing


Mentors :

  • Abhishek Mahajan

  • Isha Mukherjee

Mentees :

  • 7


In this Project We will be Studying the concept of resolutions in remote sensing, preliminaries for mathematical representation of the digital image, image enhancement, filters and transforms, feature evaluation, image classification, accuracy assessment.

We will be Working on Aerial Imagery Data obtained by MBRSC Satellite and Segmenting it Pixel-Wise into various classes/Regions by Researching and studying various Deep Learning Architectures and Algorithms involving Multi-Scale Context information.
Pre-requistes:

  • Python
  • Basic Mathematics

Tentative Timeline :

Week Number Tasks to be Completed
Week 1 COMPLETION OF BASIC PYTHON TUTORIALS
Week 2-3 UNDERSTANDING some Machine learning model related to the Project
Week 4 EVERYONE IS Required to do some TensorFlow / Pytorch coding Tutorials
Week 5-7 Development of remote sensing, characteristics of remote sensors and platforms: optical, infrared and microwave sensor systems, recent trends in active remote sensing techniques: lasers and radars, Concept of resolutions in remote sensing, preliminaries for mathematical representation of the digital image, geo-referencing, image enhancement, filters and transforms, indices and color models, feature evaluation, image classification, accuracy assessment, change detection using ML.
Week 8-10 Will train the Mentees to get familiar with the Dataset and learn to code the Data-loading and Data training part