Seasons Of Code

Digital Depth Perception    • Divyansh Srivastava   

Digital Depth Perception

Digital Depth Perception

Digital Image Perception is a simplified algorithm to have the perception of depth (or distance) based on the two different images taken at same time from different (but related) perspectives. This project is inspired by the perception of depth (or distances) by human visual system.

We have two visual receptors that capture real world image at same instant but from two related but different perspectives. We will use digital image correlation for most part of the project. We will mostly work on Google Colab.

Prerequisites:

Knowledge of functioning of a camera, how its sensor works, how images are produced and some basic knowledge of linear algebra, and should be comfortable with python language.

Proposals:

A two page proposal is mandatory in pdf form (other formats will not be entertained) which should include following points:

  1. Your motivation for this project
  2. Functioning of cameras (uses of lens, shutter, and other basic equipment present in the camera, terms like focal length, aperture, etc.)
  3. If you are given a 1000x1000 matrix A and a 4x4 matrix B, how will you find where B lies in A (write your approach and think of the scenarios where your approach may fail).
  4. Make a github profile, get yourself acquainted with its features, make a repo named Digital Depth Perception and share the link.

Tentative Project Timeline

Week Number Tasks to be Completed
Week 1 Introduction to Python, Jupyter notebook and some basic python libraries like Matplotlib and Numpy on Google Colab and algorithm analysis. Also some basic image processing based functionalities of MATLAB.
Week 2 Introduction to intensity plots and treatment of images as functions.
Week 3 Filtering of images and applying different kernels to images.
Week 4 Introduction to correlation and convolutions of images.
Week 5 Introduction to the relation between a real world scenario and the image generated by a digital camera.
Week 6 Introduction to extrinsic and intrinsic geometry of a camera and calibration of cameras.
Week 7 Image projections and introduction to different type of algorithms available online for digital distance measurement as well as their analysis.
Week 8 Creation of simple algorithm for digital distance measurement and analysis of its space and time complexity.
Week 9 Documentation of final report and analysis of real world scenarios where this algorithm can be used.