Efficiency Optimization of Dual-Axis Solar Tracking Systems Using Computational Vision Models

This paper presents an efficiency-optimized dual-axis solar tracking system leveraging computational vision techniques to enhance photovoltaic energy capture. The proposed system uses Python-based image processing to detect the brightest region in the camera frame, indicating the sun's position. The coordinates of this region are transmitted via serial communication to an Arduino Nano, which drives two servo motors to orient the solar panel along both azimuth and elevation axes, enabling real-time sun tracking. To improve tracking precision and voltage stability, four signal filtering techniques—Kalman Filter, Savitzky–Golay Filter, Particle Filter, and Moving Average Filter—are implemented and comparatively analyzed. Experimental results demonstrate that appropriate filter selection significantly enhances tracking accuracy and power output stability. This integrated system demonstrates a low-cost, intelligent solution for solar energy optimization through a combination of computer vision, embedded control, and smart communication. To monitor energy performance, a voltage sensor connected to the solar panel measures real-time output, which is read through the serial interface and displayed in a custom graphical user interface (GUI). A mathematical model estimates the charging duration based on the voltage input, and upon timer completion, the system automatically triggers an SMS alert to a registered number using the Twilio API.