Pattern recognition of lettuce varieties with machine learning / Seri' Aisyah Hassim

Seri' Aisyah , Hassim (2019) Pattern recognition of lettuce varieties with machine learning / Seri' Aisyah Hassim. Masters thesis, University Malaya.

Preview	Image (JPEG) (The Candidate’s Agreement) Download (630Kb) | Preview
Preview	PDF (Thesis M.A) Download (5Mb) | Preview

Abstract

Determination of lettuce varieties through image processing is considered as part of precision farming. Automatic classification is becoming vital for precision farming practice as it is rapidly developing with emergence of many applications for agriculture. It is a hassling process to differentiate and identify the lettuce varieties through human capabilities as it is time consuming and prone to errors in identification process. Hence, there is a need to do this assisted by a machine capability which makes it faster with greater accuracy. Application of machine learning in agricultural is still not widely applied and many phases need to be improved. Differentiation of lettuce varieties with colour or shape similarity is quite challenging. This study focuses on designing the lettuce varieties recognition by using Convolution Neural Network (CNN) in MATLAB. The neural network model consists of layers such as Convolution Layer, Normalization Layer, ReLU Layer, Fully Connected Layer, Softmax Layer, and Classification Layer. The network needs to undergo training sessions before being able to recognize the lettuce varieties. A set of data are prepared for prediction after training. The accuracy for overall classifications is 94.4% while accuracy for specific lettuce varieties of Butterhead Lettuce, Celtucelove Lettuce, Italian Lettuce, Red Coral Lettuce, Red lettuce, Red Oakleaf Lettuce and Salad Grand Rapid Lettuce were 94.7%, 99.7%, 97%, 94%, 90.7%, 98%, 87% respectively.

Actions (For repository staff only : Login required)