OPNN: Optimized Probabilistic Neural Network based Automatic Detection of Maize Plant Disease Detection
Eisha Akanksha, Neeraj Sharma, Kamal Gulati
Abstract
Agriculture is one of the important ant sectors of the Indian economy and also an important source of income for Indian people. Approximately 50% of employees worked in the Indian agricultural sector. Many crops such as rice, maize, cotton, and wheat are grown in India. Of these, maize is India's most important food crop and plays a crucial role in food production. The major problem in food production is diseases in plants. The plant diseases are identified using a leaf. Diseases in plants affect the quality and quantity of crops in agricultural production. To avoid the problem, in this study; an efficient automated diagnosis of maize plants has been developed. The proposed methodology is consists of four stages namely, preprocessing, feature extraction, classification, and segmentation. Initially, the images are converted into RGB format and the images present in the noises are removed. Then, the R band is given to the feature extraction stage. Then, the selected attributes are fed to the classifier to classify an image as normal or abnormal. For classification, an optimized probabilistic neural network (OPNN) is utilized. The PNN classifier is improved by using artificial jelly optimization (AJO) algorithm. Finally, the Northern leaf blight disease leaf images are fed to the segmentation stage to separate the affected portion of a leaf. The effectiveness of the proposed lead disease classification is investigated base on the various quality metrics namely, accuracy, sensitivity, and specificity.