Developing Leaf Identification Mobile Application For Frst Plantation Study Using Convolutional Neural Network

Phuah, Yee Ling (2023) Developing Leaf Identification Mobile Application For Frst Plantation Study Using Convolutional Neural Network. [Final Year Project Report] (Unpublished)

	PDF Phuah Yee Ling (24pgs).pdf Download (649kB)
	PDF (Please get the password by email to repository@unimas.my , or call ext: 3914 / 3942 / 3933) Phuah Yee Ling (Fulltext).pdf Restricted to Registered users only Download (4MB)

Abstract

Malaysia is one of the countries well-known for abundant biodiversity, especially in Sarawak state, which has been internationally recognised as one of the famous biological hot spots. To investigate biodiversity, many local educational institutions such as UNIMAS have offered courses in natural science. Practical training is also provided for students to enhance their learning experience and gain natural science-related knowledge. However, students of the Faculty of Resource and Technology (FRST), UNIMAS still faced difficulties to identify plant species with conventional approaches. In this project, a mobile leaf identification application named SarawakPlant is proposed to facilitate FRST students who study plant science to identify species of plants effectively and innovatively. This application works by capturing a picture of a leaf through the camera on Android mobile devices. This function is achieved by using the object-detection technique to extract characteristics such as the shape of a leaf and using this captured information to match through the pre-trained dataset. This pre-trained dataset would be taken from the Google image. The matched data with high accuracy will be returned to the users. This application is developed by using TensorFlow Lite and Android Studio. Moreover, supervised learning is applied in this leaf identification mobile application to integrate with the pre-trained dataset. Lastly, this system aims to engage students and assist them to have higher accuracy in plant species identification.

Actions (For repository members only: login required)

View Item