Correlation estimation between nitrogen and bean plant colors

Abstract

Processing digital images has a lot of applications in different sciences such as medicine, industry and agriculture. One of the uses of digital images is in agriculture industry, for instance digital images could be used in providing the nitrogen of the plant. This research aims to estimate the correlation between the amount of nitrogen in bean plants and its color parameters. For this goal an algorithm is proposed in this paper. First beans images are provided and some preprocessing operations such as resizing, noise removing are performed. Second, RGB color space is converted to HSV color space. Finally correlation between plant color and nitrogen is estimated using regression equation. Implementation results show that there is high, strong, and positive correlation between the color features and the amount of nitrogen in the tissue of the bean. Among RGB, green color has the highest correlation with nitrogen compared to other colors, it is about 0.62. In addition, we consider the combination of three colors for our estimations. Our study show that (G/G+R) has the highest correlations with nitrogen in comparison to other equations and results. It is about 0.81.