Overview / Usage

This project aims to detect activity in pedestrians, in drone videos. The resulting work/research will result in being used in areas such as city or event surveillance, crowd management and border patrolling.

Much of the security concerns of the society will be addressed through this resulting research work.

Methodology / Approach

I'm trying to address the given problems using Deep Learning.

I've started by, firstly working on Tensor Flow, and then moved on to seeing which of the detection frameworks will give the best standoff between performance and accuracy. After working on pedestrian detection and analyzing these performances, I can move on to activity detection.

Technologies Used

• Tensor Flow and PyTorch, Python, OpenCV were used in this project.
• SSD and YOLO were used as detection algorithms.

​

​