Enhancement Of Segmentation And Feature Fusion For Apple Disease Classification

Diseases in fruit cause serious problem in economic loss and production in agricultural industry. The quality and yield of the fruits can be degraded too much with the presence of the diseases in the fruit. It is critical to detect the diseases present in fruit and monitor fruit's health. Advanced computational methods have been used in agricultural applications, such as disease detection, classification and grading to help agriculturists in common farm tasks with more precision, efficiency, productivity and cost reduction. An approach for the apple disease classification with enhancement of defect segmentation and fusion of color, texture and shape based features is used to classify the apple into two types, healthy and defected with an accuracy of 96% using Histogram of Oriented Gradients feature descriptor and Bagged Decision Trees classifier.