Enhanced CNN and PCA Features Based Leaf Disease Classification

  • Anitha K., Dr. Srinivasan S.

Abstract

The invention of agriculture for sustainable living is one of the great revolutions in the history of human. It significantly changed the human society in food production and domestication and played a major role in population increase and biological changes.. The proposed research, timely decision on agriculture produce and disease control is done with design and development of disease diagnosis system using android application for image processing and machine learning techniques, also irrigation parameters are monitor to control water utility of agricultural field. This proposed work has further wavelet based statistical features are used to increase the accuracy for diseases classification so that the yield can be increased to a great extent by taking appropriate control measures. In this research work wavelet decomposition with PCA (WPCA) based statistical features are extracted resulting in a set of total 11 features. All these features are applied as input to artificial neural network for classification of diseases occurring in plant parts. The combined result of  WPCA based statistical features shows the accuracy of 98.280 for flower, fruit and leaf respectively, This work also used the Enhanced convolution neural network  (EDCNN) to train with  all the features provides the accuracy of 98.240 for flower, fruit and leaf images respectively. Hence, the proposed work WPCA features with use of enhanced artificial neural network turns out to be a better method in order to diagnose the pathological issues for crop with different plant parts based on android application for image processing and machine learning techniques.

Published
2020-10-03
How to Cite
Anitha K., Dr. Srinivasan S. (2020). Enhanced CNN and PCA Features Based Leaf Disease Classification . International Journal of Advanced Science and Technology, 29(04), 10210 -. Retrieved from http://sersc.org/journals/index.php/IJAST/article/view/33059