Non-destructive Estimation of Spinach Leaf Area: Image Processing and Artificial Neural Network Based Approach

Abstract

Leaf area (LA) measurement provides valuable key information in understanding the growth and physiology of a plant. Simple, accurate and non-destructive methods are inevitable for leaf area estimation. These methods are important for physiological and agronomic studies. However, the major limitations of existing leaf area measurement techniques are destructive in nature and time consuming. Therefore, the objective of the present work is to develop ANN and linear regression models along with image processing techniques to estimate spinach leaf area making use of leaf width (LW) and length (LL) and comparison of developed models performance based on the statistical parameters. The spinach leaves were grown under different nitrogen fertilizer doses (0, 50, 100, 150, 200, 250, 300, 350 and 400 kg N/ha). The morphological parameters length (LL), width (LW) and area (LA) of leaves were measured using an image-processing software. The performance LA= -0.66+0.64 (LL × LW) (R2 = 0.98, RMSE = 3.25 cm2) equation was better than the other linear models. The performance of the ANN model (R2 = 0.99, RMSE = 3.10 cm2) was better than all other linear models. Therefore, developed models along with image processing techniques can be used as a non-destructive technique for estimation of spinach leaf area.