Degradation Analysis of Banana Using Comparative Approach

Abstract

In recent times, the considerations of improving the product and food quality are turned into a big concern. Continuous uses of chemical, formalin and so many preservatives on food and agriculture are now a burning question among the mass people and becomes a headache for the government to ensure food freshness security to among the mass people. Hence this turns us to come up with this idea of food freshness and degradation analysis. Primarily we have started our research on Banana fruit degradation analysis to ensure banana’s freshness. Comparative research and comparative degradation analysis of banana can be a good solution to ensure food quality and freshness of banana. The aim of this project is to research degradation analysis of Banana. The type of Banana that is used in this research is called Sagor (AAA genome [a]Musa Sapientum). Two different methods Convolutional Neural Network (CNN) and CONTOUR are used for more comparative study to reflect on degradation analysis. Changes in color shape and black spots of 67 fresh samples of Bananas are observed for seven days. The samples are put at room temperature for seven days. The changes of each of the banana sample are recorded from day 1 to day 7. For CONTOUR model banana images are captured from two different angles to investigate features for degradation analysis and images are captured using normal resolution smartphone camera and merge them into one image to be used as the test set. On the other hand for CNN model single 134(One hundred and thirty four) one side banana images are taken and after that by using augmentation the datasets are increased into 300 banana images which are used as test set for CNN model. These two comparative approaches provide two magnificent outputs with pleasant accuracy where CNN gives 90% accuracy and CONTOUR provides 60% accuracy. In the future further algorithmic development will be implemented including AI for better approximation.