Loading...

Application of image processing and linear regression models for estimation of nitrogen content of tomato leaves

DOI: 10.31830/2348-7542.2019.051    | Article Id: 016 | Page : 345-352
Citation :- Application of image processing and linear regression models for estimation of nitrogen content of tomato leaves. Res. Crop. 20: 345-352
Asma Kisalaei, Vali Rasooli Sharabiani, Ebrahim Taghinezhad vrasooli@uma.ac.ir
Address : 1Department of Biosystem Engineering Faculty of Agriculture and Natural Resources, University of Mohaghegh Ardabili, Ardabil, Iran; 2Department of Agricultural Machinery, Moghan College of Agriculture and Natural Resources, University of Mohaghegh Ardabili, Ardabil, Iran

Abstract

Application of variable rate technology (VRT) of fertilizer on the farm is a major pillar of production accurate management that increases the efficiency of fertilization and reduces the pollution of environment. Tomato is an important valuable vegetable that is in second place after potato in terms of economic and food in the world. This study formulates a mathematical relationship correlating the nitrogen values of tomato using an image processing to the common methods during the growing season. The nitrogen content of leaves was measured by two common methods including chlorophyll meter (SPAD) and Kjeldahl test. The measurements were performed in two stages, once before fertilization and the other one week after fertilization. In addition, with cutting of plant leaves, it was attempted to take pictures using a digital camera in the chamber that was controlled in terms of light. All pictures were transferred to the space on MATLAB R2013 software and average colour of R, G, B and 11 colour models were extracted from RGB images. It was found that a linear relationship existed between colour parameter (G-band) and SPAD (R2=0.96) as well as between G-band and nitrogen content (Kjeldahl test) (R2=0.97). Results of this research can be utilized for quick and simple determination of SPAD and nitrogen values and eliminating the need for conventional physical-chemical methods.

Keywords

Image processing  modelling and estimation  neural networks  nitrogen content  tomato.

References

Global Footprints