Authors:
K. N. V. Satyanarayana, Adireddy Jahnavi Vagdevi, Dadi Pavithra Sree, Boddu Mouni, Chintapalli Sri Satya Kanaka Rajesh, Dinka Lale

Addresses:
Department of Electronics and Communication Engineering, Sagi Ramakrishnam Raju Engineering College, Bhimavaram, Andhra Pradesh, India. Faculty of Electrical Engineering and Applied Computing, University of Dubrovnik, Dubrovnik, Dubrovnik-Neretva County, Croatia. 

Abstract:

Most accidents that involve falls from heights, such as on terraces and balconies, cause severe injuries. They occur in both residential and industrial environments. The safety systems now available primarily focus on detecting falls after they occur, typically using wearable sensors or image-based classification. Still, they do not provide complete preventive measures. This paper presents a real-time fall detection and protection system that is based on pose estimation and embedded actuation. The system uses OpenCV for capturing video and MediaPipe Pose to identify landmarks on the human body, particularly key points on the hip and ankle. A dual-boundary spatial model detects when a person is closer to a dangerous edge zone when their feet are above ground level. This system distinguishes a person climbing from a static position using temporal motion analysis. This is judged by checking whether the risk conditions are satisfied for at least six consecutive frames. After that, an alert signal is transmitted to an ESP32 microcontroller, triggering a buzzer alarm and activating a servo-driven safety net. In general, the system provides a practical approach to managing safety in risky environments.

Keywords: Fall Detection; Pose Estimation; Embedded Systems; Human Activity Recognition; Pre-Fall Risk Detection; Image-Based Classification; Human Pose Estimation; Embedded Safety Mechanism.

Received on: 24/08/2024, Revised on: 01/11/2024, Accepted on: 12/01/2025, Published on: 03/06/2026

DOI: 10.69888/FTSES.2026.000675

FMDB Transactions on Sustainable Energy Sequence, 2026 Vol. 4 No. 1, Pages: 11-21

• Views : 21
• Downloads : 7

Download PDF