A NOVEL APPROACH OF DAMAGED PADDY LEAF DETECTION TECHNIQUE USING BACK PROPOGATIONNEURAL NETWORK (BPNN)

The environmental influences (i.e. soil, weather) on paddy cultivation make a significant contribution to the worldwide rice production rate. Effective management of paddy diseases and pests is, however, the next major influence on growing production. Every year, farmers lose an approximate average of 37% of their rice paddy leaves to pests and diseases. In order to prevent any kind of disaster that can be caused by diseases, it is crucial that farmers recognize the condition of their paddy well in advance before it is too late. Accurate diagnosis and timely resolution of paddy disease is therefore a critical component of the management of rice production aimed at improved productivity leading to higher income. As a result, farmers rely on their experience and intuition to decide on the identification of paddy leaf diseases and their treatment in most cases. If the symptoms are not adequately treated, the production may not turn out as expected, using sufficient quantities of fertilisers directed by agricultural specialists. Traditionally, the damage diagnosis of various paddy leaves has been conducted manually. The proposed work incorporates the methodology called k-means clustering and back propagation neural network for classification, the multi-stage machine learning method. There are two types of NN based on the method of finding out, the place output values are known in the past (back propagation algorithm) can be monitored and the place output values (clustering) cannot be known unmonitored. The excellent performance of the proposed work has been obtained through experimental findings.