Visual Computing

POSE ESTIMATION USING RASPBERRY PI

Faculty Mentors:

• Arthi B
• Aruna M

Project Members:

• Akul Abrol (RA2211003010029)
• Paras Vichoray (RA2211003010040)
• Prashuk Jain (RA2211003010027)
• Tushar Ahlawat (RA2211003010056)

PROJECT OVERVIEW:

Pose estimation is a computer vision technique that classifies the different body parts and joints of a human body and tracks their movement. This is achieved by assigning different key-points or ‘landmarks’ to the parts and joints of the body, and thus continuously tracking their real-time location through images or video feed while providing the user with the coordinates of those landmarks.

This project made use of Raspberry Pi, more about which will be explained later. The project was made in Python coding language, and the OpenCV and MediaPipe libraries act as the most important parts. The Raspberry Pi camera module was used to get live video input and various other pre-recorded videos were also used to test the project out.

Project title: Human detection in synergy with OSCC detection

• Hardware used (As per equipment in VC lab): Microscope and Mi 360° Camera, Oculus Quest 2
• Software used (Frontend & Backend): Python, Flask
• Problem statement: creating a synergy of OSCC detection and Human detection using an NLP.

Project title: SRM Crowd Detection and Security System

• Hardware used (As per equipment in VC lab): Wifi 360° camera
• Software used (Frontend & Backend): Python, Opencv, Tensorflow, CNN algorithm, flask

Problem statement:

Crowd detection and management in college settings present multifaceted challenges that extend beyond conventional security concerns. Large gatherings of students, faculty, and visitors are commonplace in academic institutions, leading to issues related to safety, resource optimization, and campus experience. Current systems and approaches may not fully address the unique needs and complexities of college campuses. Identifying and addressing these open issues is essential to create safer, more efficient, and more inclusive learning environments.

Project title: Lane Detection System

• Hardware used (As per equipment in VC lab): Microscope and Mi 360 ° Camera.
• Software used (Frontend & Backend): Python, Flask
• Problem statement: Creating a lane detection system for self-driving cars.

Project title: Computer Vision-Based Autonomous Parking Space Detection.

• Hardware used (As per equipment in VC lab): Raspberry Pi, Camera
• Software used (Frontend & Backend): Python, OpenCV
• Problem statement: Parking management in our college has become increasingly challenging. Finding an available parking space quickly and efficiently has become a significant concern. To address this issue, we are developing an advanced Parking Space Detection System that can accurately and intelligently detect available parking spaces.

Project title: Realtime Wildlife Conservation System

• Hardware used (As per equipment in VC lab): Raspberry Pi, Camera
• Software used (Frontend & Backend): Python, OpenCV, Tensorflow, Keras, Megadetector, YOLOv5
• Problem statement: The conservation of wildlife is a critical concern worldwide, with a pressing need for more efficient and accurate monitoring systems. Current wildlife monitoring methods often suffer from limitations in terms of real-time capabilities, accuracy, and scalability. Traditional approaches rely heavily on manual observation and are unable to provide timely information for immediate action. There is a clear need for a robust and real-time wildlife target detection system that addresses these issues.