Abstract:
Throughout the world, papaya seems to be very familiar both as a vegetable and as a fruit, depending on the stage of maturity level. If we look over other countries, we can find that papaya has a large amount of values because of its health benefits & also protect against a number of health conditions which is proved by medical science from many years before. A tropical fruit with excellent nutritional & therapeutic appraisal due to its abundant origin of vitamins A & C is the Papaya (Carica Papaya). South Mexico and Costa Rica are where papaya harvesting first began. Fruit output is expected to be 6 million metric tons annually throughout the planet. With an annual output of approximately 3 million tons, India is the globe's top producer of papaya. Brazil, Mexico, Nigeria, Indonesia, China, Peru, Thailand, & the Filipinos all additional top inventor. It’s very much sensitive to frost, strong winds and water stagnation and also rotten very fast. The main objective of this study would present non-intrusive classification system for papaya fruit perfection stages. For this project we want to build a system using Machine Learning also known as python programming language or Deep learning which can perfectly detect papaya maturity and its classification. When developing we learn about different types of machine learning algorithms and techniques like GoogleNet, AlexNet, VggNet, Resnet, LeeNet, CNN, KNN, Linear Regression, Logistic Regression, SVM, Random Forest, K-Means cluster, Decision Tree, Naive Bayes etc. to understand the whole thing and to complete the project also. The main purpose of this study is to show the maturity level & its classification perfectly. For image classification, CNN model performs best then other models so we decide to build a sequential model for this project. In this case to detect our algorithm, we need to use a huge number of image dataset for training, testing & validation purpose. The model we choose & offered for detection successfully found the outcome. Our preliminary findings demonstrate that the methodology we provided achieved 100% consistency.
