Detection of Disease in Banana Fruit using Gabor Based Binary Patterns with Convolution Recurrent Neural Network

Abstract

Banana plant disease classification is an application which supports farmers by making easier to analyse, detect and control plant pathogens. In order to protect the crops with the feasible cost, banana crop infection symptoms need to be identified and treated at the initial stage. This can be analysed and bifurcated through the computer vision system which uses interpretation of information by image processing techniques. Banana, fruit of the genus Musa, of the family Musaceae, one of the most important fruit crops of the world. Determining banana’s disease detection stages is becoming an essential requirement for standardizing the quality of commercial bananas. So this paper proposes the novel feature extraction technique in extracting the stages of disease detection of banana using Gabor based binary patterns with convolution recurrent neural network. The collection of fine-grained features of image on basis of mechanism which is driven by data also it provides the phases of disease affecting in banana fruit. For variation of symptoms which is resulted has been assisted for variations between subsequent groups of banana for disease affecting. The simulation results has been taken from banana image of 17,312 which shows various stages of disease growing in banana fruit which shows that the proposed neural network obtains enhanced accuracy on basis of computer vision for both classification in rough- and fine-grained for disease affecting stages of banana which cause severe impacts.