Herb leaves recognition

Aziz, Nur Syafiqah Farhah (2017) Herb leaves recognition. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

	Text (24 Pages) Herb leaves recognition.pdf - Submitted Version Download (334kB)
	Text (Full text) Herb leaves recognition.pdf - Submitted Version Restricted to Registered users only Download (1MB)

Abstract

Herbs are any plants used for food, flavoring, medicine, or fragrances for their savory or aromatic properties. Culinary use typically distinguishes herbs from spices. Herbs refers to the leafy green or flowering parts of plant (either fresh or dried), while spices are produced from other parts of the plant (usually dried), including seeds, bark, roots and fruits. The purpose of this project is to develop the herbs’ leaves type recognition system. The known of herbs become less conscious among people nowadays, by using this herb’s leaves type recognition it will ease the manpower to recognize the leaf type. Recognition of plants is a challenging computer vision problem that requires dealing with irregular shapes and textures with high intra class variability. The main leaves features considered in this project are leaves major length, minor length, and area. To complete this project, the method decided to use is image processing method which is to detect the herbs’ leaves type by using MATLAB R2012b software. This project used K-Nearest Neighbors classifier (KNN) and Linear Discriminant Analysis classifier (LDA) to classify the dataset. The purpose of using the KNN classifier is because this classifier is robust to noisy training data. Thus, it is more effective for large training data. The purposed of LDA method is to improve the discriminative ability of projected samples. The dataset for this project are divided into two which are for testing set and training set, 60 and 180 dataset respectively. The images captured are grouped according to their types. The expected outcomes for this project is the system can recognize the herb leaves type correctly according to their classes and groups.

Actions (login required)

View Item