Visual SLAM and 3D reconstruction
Simultaneous localization and mapping (SLAM) is the computational problem of constructing or updating a map of an unknown environment while simultaneously keeping track of an agent’s location within it. It is a widely popular algorithm for mobile robot localization and 3D reconstruction. Current state of art SLAM systems require high computation and are not suitable to use for real time application in low resource devices. Some available low resource real time SLAM algorithms are not robust enough to tackle the very dynamic nature of the real environment.
This project focuses on developing models and algorithms which can perform SLAM in real time on low resource devices such as raspberry pi and jetson nano.
You will be working in a research group with Dr. Ajad Chhatkuli and Rashik Shrestha.
- Good programming skills in C++ and Python
- Experience with deep learning frameworks like PyTorch and Tensorflow
- Excellent results in subjects such as linear algebra, computer vision, machine learning, image processing and statistics.
- Strong communication skills for methodology and results of experiments.
- Knowledge about SLAM and multi view geometry
- Solid scientific writing abilities in English, as well as the ability to visualize experimental data with graphs and figures.
- Knowledge of git version control
- Enhancing available SLAM frameworks .
- Implement or modify current pipelines and basic software platforms to acquire appropriate data.
- Identify and highlight critical issues with existing models.
- Disseminate research findings to wider populations through publishing in international conferences and journals.
16 hours per week
Minimum Required Qualifications
Studying Bachelor’s in Computing Sciences & Engineering related fields (Exceptions are possible if you have exposure to Machine Learning)
- Opportunity to work in a research environment
- Get trained on how to read and write papers
- Opportunity to get close mentorship of NAAMII researchers