Leaf area estimation based on ANFIS using embedded system and PV panel

Leaf area is one of the important parameters for plant canopy development. It is used as an indicator closely related to plant growth in several studies on plant production. However, most leaf area meters used today are costly and rely on human observations. This situation may be limiting for researchers in terms of having proper leaf area measuring devices. The reliance on humanfocused measurements leads to human errors. Digital scanners and cameras, digital image processing-based estimation methods, paper weighing, grid counting, regression equations, width and height correlation models, planimeters, laser optics, and handheld scanners can be used to determine leaf area. However, some of these methods are expensive and unnecessary for simple studies. Therefore, this study aims to design and implement an embedded system with a simpler, cheaper alternative to the currently used methods and devices, minimizing human errors. The proposed embedded system serves as a tool for measuring leaf area using a photovoltaic panel (PV) and an Adaptive Neuro-Fuzzy Inference System (ANFIS). In the study, geometric shapes with known areas are used as the learning data, and real plant leaves with known areas are used in the testing process. As a result, the prediction made by ANFIS is observed to have an accuracy of R2 = 0.99.