Automatic leaf color level determination for need based fertilizer using fuzzy logic on mobile application: A model for soybean leaves

Abstract

Detecting plant nutrient deficiencies and evaluating fertilizer program are done by leaf tissue analysis. Unfortunately, this quantitative method is quite expensive and time consuming for traditional farmers due to its laboratory procedure. In this research, an automatic and non-destructive method based on digital image for soybean leaf color level determination was developed. Color level status is used to determine the fertilizer dose based on crops current need. The color level was adopted from 4-panel Leaf Color Chart (LCC) and a fuzzy logic model was applied to capture the leaf color gradation. Therefore, the leaf color status is not restricted only in 4 categories, but gradually change from light yellow up to dark green. Using this mechanism the N fertilizer dose will also gradually adjust. Hence, the N fertilizer could be used efficiently and in the same time prevent the environment from negative effects of fertilizer overuse. The method was embedded in a mobile application to facilitate real time field application. Hence, detection of soybean nutrient deficiencies and fertilizer program evaluation will need less time and low cost. From the field test, it was known that the mobile application could determine the soybean color level correctly.