Maximizing energy efficiency in a solar tracking system using iot-based image processing

In this thesis, an image processing-based solar tracking system was developed and integrated with the IoT platform ThingsBoard. The main objective was to detect the sun's azimuth and elevation angles from a fixed camera through image frames and to utilize this information in both simulation and potential physical applications. OpenCV was used for image processing. Through HSV-based thresholding and contour analysis, the bright region corresponding to the sun was isolated, and its center coordinates were calculated. These coordinates were converted into azimuth and elevation angles and transmitted to the ThingsBoard platform via the MQTT protocol. On the platform, the data was visualized in real time using time series charts, radial indicators, and rule-based alarm mechanisms for system monitoring. Additionally, the Hottel atmospheric model was implemented to perform theoretical solar irradiance estimations based on the calculated elevation angles. These estimations were also integrated into the ThingsBoard platform, allowing for a time-dependent comparison between visually detected sun positions and theoretical irradiance values, thus validating the system's accuracy and real-world reliability. To demonstrate the feasibility of physical integration, a Python-based servo motor control structure was prepared. The system was designed to transform the received azimuth angle into physical movement, enabling motor rotation via PWM signals. Although not implemented physically, the control logic was built to be ready for hardware deployment. Overall, the developed system presents a low-cost, modular, and scalable solution combining real-time data monitoring, automation, and theoretical modeling—making it suitable not only as an academic prototype but also for real- world solar energy optimization applications.