Nutrition information extraction of the cucumber leaves in the greenhouse based on computer vision technology

The method to diagnose the growth conditions of cucumber in greenhouse was studied with the machine vision. The images of the cucumber leaves were taken under the sunlight condition, then the red, green, blue (RGB), their relative ratios (r, g, b), and the hues of the images were calculated. And the correlations among nitrogen, phosphorus and water content of the leaves and their color parameters were analyzed using the RGB and the HSI model. The result shows that there are high linear correlations between the nitrogen content and the green weight, and between the nitrogen content and the hue so that two parameters could be used as the indices of the growth for the fast diagnosis using machine vision. Whereas the other color weights had not so high correlation with the nitrogen. It was observed that the color weights did not have obvious correlations with the phosphorus and water content. Additionally, it was found that different light conditions could have an effect on the linear relationship between the nitrogen content and the green weight or hue. So the method needs to be improved for higher precision of the linear regression through further experiments under the artificial light source and the system calibration.