Detection and Classification of Crop Disease Using Three-channel CNN with ResNet Architecture

Authors

  • Sagana C,Sangeetha M , Manjula Devi R, Harshavarthini S, Gokul K, Kavin M

Abstract

In agriculture, crop disease affects the growth and yield of the crop. Detection of disease is the most crucial part of agriculture. Manual detection and classification of the disease are more difficult and time-consuming, it requires more time, expert knowledge and not suitable for large fields. The three-channel convolutional neural network(TCCNN)  is used for the detection of leaf disease. Colour information of leaf is used for plant leaf disease detection and classification. Each channel of TCCNN is fed by three color components of RGB diseased leaf image. Two different types of crops are used for disease detection. The dataset contains various diseased leaf images and healthy leaves of tomato and maize. Along with TCCNN, Resnet architecture is used to train the network deeper and this model requires minimal image preprocessing process. Experimental results show that Three-channel CNN with Resnet architecture achieved better accuracy.

Published

2020-11-01

Issue

Section

Articles