A Cloud Computing-Enabled ESP32-CAM System for Real-Time Object Recognition with Feedback

This paper presents a hybrid IoT architecture that integrates edge computing with cloud processing for real-time, low-cost object recognition. It utilizes an ESP32-CAM, a popular microcontroller with WiFi capabilities integrated with a camera sensor, to capture images and transmit them via Message Queuing Telemetry Transport (MQTT) to AWS, where a Single Shot MultiBox Detector (SSD) model trained on the Common Objects in Context (COCO) dataset processes them for object detection. The system provides in-circuit feedback by controlling General Purpose Input/Output (GPIO) pins, enabling immediate action execution. Performance evaluations show an object detection accuracy of $81.25-83.6 \%$ with an average latency of 5.1 seconds. The study highlights scalability, security, and cost-efficiency, with future improvements targeting reduced latency, enhanced security, and expanded industrial applications.