A.Processing of the image: The image is taken from a High quality camera and the image acquisition is done in this step. A focused                                                             image of the leaf is cropped manually.
B.Remove noise and unwanted features: The noise and unnecessary features are removed using a Gaussian filter.
C.Extract the RGB values and texture indexes: The RGB values ranges between 0 to 255 and the texture index indicates the water                                                                                                            content of the leaf
D.Create a dataset from the extracted features: The RGB values and other features are extracted and are stored in the database for                                                                                                             further use 
After creating a data set for minimum 100 rows, the next step is to use it for the machine learning model.In this work we are using an unsupervised machine learning i.e clustering technique because we cannot label the classes as soon as the features are extracted. The machine will provide a better accuracy if the data is unlabeled and are in the form of clusters. This technique will put each of the samples in a particular cluster. This cluster is nothing but the deficiency or a class of deficiency of the sampled leaf. This makes the identification of the text image easier. The following flowchart gives a basic idea of the machine learning process.