Using Machine Vision for Indoor Plant Recognition in a Mobile Application

This master’s thesis focuses on developing a mobile application that facilitates the care of houseplants. The project is created as an effort to create a free alternative to commercial applications on the market. The main goal of the thesis is to create a functional prototype of an application, that can recognize common indoor plants from an image and provide care information. Another feature mentioned in this work is the ability to set reminders or add plants to a collection within the app. At the time of writing this thesis, there are no mobile applications on the market that meet the specified requirements and are available completely free of charge. The work is divided into several parts. Initially, the reader is introduced to business analysis, that discusses the problem being addressed, the market competition, and the requirements for the application. Subsequently, the technical solution is discussed, focusing on the appropriate selection of technologies. The next section is dedicated to data collection, which is utilized across the application. After assembling the data set for training the classification model, the training of the neural network capable of correctly classifying commonly occurring indoor plants is addressed. The fifth part of the work focuses on the development of a cross-platform application that meets the set goals of the thesis.