Projects
Monoculor Visual Odometry
Visual odometry(VO) is the process of determining the position and orientation of a robot by analyzing the associated camera images. The project is designed to estimate the motion of calibrated camera mounted over a mobile platform. Motion is estimated by computing the feature points of the image and determining the relative translation and rotation of the images.
Pipeline includes:
1 Image sequences
2 Feature Detection
3 Feature Matching (or Tracking)
4 Motion Estimation (2-D-to-2-D,3-D-to-3-D, 3-D-to-2-D)
Autonomous Traffic Analysis And Monitoring Systems
Aimed to develop Automatic traffic analysis and monitoring pipeline.
Implemented Object Detection, Tracking, concepts of Computer
Vision and Photogrammetry to estimate the motion of a vehicle/pedestrians.
The Pipeline can be used to calculate:
1. Vehicle speed and acceleration profile at every instance over the period.
2. Trajectory of vehicles in the region.
3. Classification of the vehicle types.
4. Calculation of PCU (Passenger car unit).
5. Gap analysis.
The pipeline is robust and shows 92% test accuracy for vehicle classes and 80% for pedestrians and bicycles.
Currently working for the deployment in the traffic department using Jetson Nano.
Autonomous Drone For Martian Environment
Designed Mars aerial system (aerial vehicle) which shall be fully equipped and mission ready for operation on Mars.
The challenge was organized by Mars Society South Asia's International Planetary Aerial Systems Challenge (IPAS Challenge) 2021.
​
Uplift Capacity of Earth Anchors
In Progress: Currently working on inclined anchors and double layer soils
​
