Approximation techniques for apple disease detection and prediction using computer enabled technologies: A review

Abstract

It has been observed that Nutrition is the main driver of human living beings, hence agriculture and its crop yield become the major economic growth driver of the country. There are lot of reasons available in literature which affect the yield of the crops. It is found out that plant disorders/diseases play a vital role in affecting the yield of a crop. This work is an attempt to analyse the contribution of computer assisted technologies in the detection and prediction of apple crop diseases. Here it is observed that lot of authors provide different solutions for detection and prediction of crop diseases at the growing stage of the plant by using machine learning, deep learning and domain based techniques. Out of the present works, a few authors have emphasized harvested crop infection. These works are here by analysed for their respective strengths and weaknesses in terms of their performance measurements like accuracy, Precision, Recall, F1 Score, speed and error rate for the disease detection. It is further observed that out of the various available techniques, machine learning has got its own limitation for detection of apple crop disease and similar is the case with deep learning techniques. It is argued that convolutional neural network models in association with deep learning technique has got greater potential for producing satisfactory result (in terms of accuracy, Precision, Recall, F1 Score) when applied to real-time image data, such as 3D real-time image. Another interesting dimension attached to this work points towards the exploration of the biotic and abiotic factors responsible for the incidence of apple plant disease and subsequent detection.