An Employed of PNN algorithm in Smart Farming System for Early Plant Disease Detection and Diagnosis

Abstract

Crops are vulnerable to adverse weather conditions, which lead to lower agricultural yields. It affects the agriculture of the world. In addition, the condition proves to be very serious when the plants are infected by patients. Current agricultural methods and structures are here to recognize plants and prevent them from the effects of various infections. In this study, some strategies and methods are developed in the preparation phase and highlighted in the extraction phase to establish the accuracy of a qualified home to distinguish algorithm -dependent PNN-AI pepper leaf disease.The basic commitment of this work is to create a new database of pepper plants named (DiyalaPepper) with 166 images with some difficulties, such as obtaining images of pepper diseases, in better imaging places and conditions where recording under the same conditions is only possible in the laboratory. It is a test to see if the photos were taken under the same conditions in a normal room due to the wild climate. Furthermore, the data must be separated by 30% during testing and 70% during production. The initial Experimental evaluation of the proposed approach achieved the best accuracy =92.10 during  testing phase and  accuracy (100%) during the preparation phase.