Smart Child Safety and Ferocious Dog Behavior Detection System using YOLO and IoT

Internet Of Things

Child safety and protection from animal attacks are important concerns in public and residential environments. This paper presents a proposed smart monitoring system that combines deep-learning-based computer vision with Internet of Things (IoT) technologies to support early warning and emergency communication. The system focuses on two related tasks: identifying dogs and classifying potentially aggressive dog behaviour from camera images, and monitoring a child’s safety using wearable sensors and location tracking. A YOLO (You Only Look Once) object-detection model is proposed for analysing video frames, while OpenCV supports camera input and image processing. An ESP32 microcontroller collects data from the heartbeat sensor, SpO2 sensor and galvanic skin response (GSR) sensor. A GPS module provides location coordinates, and an SOS push button allows a child or caregiver to raise a manual alert. When the vision model detects a suspected animal threat, or when sensor readings or the SOS input indicate a possible emergency, the system can activate a local buzzer and send a notification with available location information to a parent or guardian through a web or cloud interface. Python, Flask and ThingSpeak are considered for the software and monitoring layer. The proposed architecture is intended to reduce dependence on continuous manual supervision and improve awareness of potentially dangerous situations. The system’s usefulness will depend on the quality and diversity of the training dataset, camera conditions, sensor calibration, network availability and suitable alert thresholds. Physiological readings are treated only as indicators that may require attention, not as proof of fear or kidnapping. Likewise, image-based classification is intended to provide an alert for human review rather than a definitive judgement of an animal’s intent. Any physical deterrent or electrical mechanism is excluded from the recommended safety design unless independently assessed and certified by qualified safety professionals. The proposed system can be explored for schools, parks, residential areas and other child-safety applications.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.