Intelligent Global Maximum Power Point Tracking using Neighboring Pixel Algorithm and classification and regression trees assisted Virtual Imaging under partial shading conditions

Partial shading is one of the most important issues associated with photovoltaic systems, leading to formation of several local maxima in the power-voltage curve and thus reducing the efficiency of energy production. Traditional MPPT methods find it difficult to determine the Global Maximum Power Point in partial shading conditions. In this work, a novel intelligent Virtual Imaging Maximum Power Point Tracking method for the real-time detection of Global Maximum Power Point tracking under partial shading conditions is applied. The applied technique consists of digital image processing techniques, Neighboring Pixel Algorithm, and Classification and Regression Tree based shading classification techniques. Real-time images of the PV array are taken using a camera, and based on the acquired images, module classification into non-shaded, partially shaded, and fully shaded modules is performed, and accordingly, the reference voltage is adjusted to detect the Global Maximum Power Point tracking. The proposed Virtual Imaging Maximum Power Point Tracking framework was implemented on a Raspberry Pi 4 platform using MATLAB R2024b for real-time image processing and adaptive Maximum Power Point Tracking control. The effectiveness of the proposed framework is verified for a 3 × 2 PV array with vertical, horizontal, and diagonal shading conditions. Experimental results demonstrate a tracking efficiency of up to 99.9% with a response time of 5–10 ms. The proposed method accurately identifies shading patterns and rapidly converges to the Global Maximum Power Point tracking. A comparative study is performed against Perturb and Observe, Incremental Conductance, and artificial intelligence-based Global Maximum Power Point tracking methods. The results show superior Global Maximum Power Point detection accuracy, faster tracking, and lower computational complexity. The proposed Virtual Imaging Maximum Power Point Tracking framework offers a low-cost, scalable, and practical solution for rooftop and utility-scale photovoltaic installations operating under partial shading conditions.