Open AccessOptimal Deep Learning Model to Identify the Development of Pomegranate Fruit in Farms
Author(s) -
Sripad Joshi,
Sandeep Kumar Panda,
Sigal Ar
Publication year - 2020
Publication title -
international journal of innovative technology and exploring engineering
Language(s) - English
Resource type - Journals
ISSN - 2278-3075
DOI - 10.35940/ijitee.c8762.019320
Subject(s) - ripeness , economic shortage , artificial intelligence , artificial neural network , class (philosophy) , computer science , stage (stratigraphy) , receiver operating characteristic , machine learning , process (computing) , pattern recognition (psychology) , agricultural engineering , sensitivity (control systems) , mathematics , horticulture , ripening , engineering , biology , paleontology , linguistics , philosophy , government (linguistics) , electronic engineering , operating system
In this study, estimating the maturing condition in gardens helps to enhance the process of post-harvesting. Collecting fruits on the basis of their developmental stage will minimize storage costs and maximize market value. Additionally, estimated ripeness of the fruits can be more useful for indicators for detecting water shortage and to determine the water used during irrigation. The purpose of the study is to develop the new direction of technology to detect the ripeness stage between two classes: ripe and unripe. We employ deep Neural Network (DNN) classifiers for the prediction of ripe and unripe class. The results of our proposed classifiers give the sensitivity 96.2%, specificity 94.2% with accuracy of results 94.5%, over a dataset of 200 images of each class. The ROC (receiver operating characteristic) area values curve close to 0.98 in all-class during training. We believe this is a notable performance that allows a suitable non-intrusive maturing prediction that will enhance cultivation techniques.