MARIJUANA DRUG DETECTION AND CLASSIFICATION USING MACHINE LEARNING APPROACH

Adugna, Assefa Woleli

dc.contributor.author	Adugna, Assefa Woleli
dc.date.accessioned	2021-09-21T10:48:19Z
dc.date.available	2021-09-21T10:48:19Z
dc.date.issued	2021-07
dc.identifier.uri	http://ir.bdu.edu.et/handle/123456789/12618
dc.description.abstract	This study investigated to detect and classified Marijuana drug (mind alert substance) from other bush and crop plants. Data were collected from farm lands and individual compounds. It’s properly segregated and all dataset images are identified and are renamed. Amhara police commission is volunteer and cooperate to this study. The commission has written a guarantee letter to collect those drug image for only research purpose therefore, 2000 images are gathered in Bahir Dar city, Ethiopia. To improve the collected image data by selection and enhancing the important image features using resize, grayscale, denoise, contrast enhancement, segmentation and augmentation which are necessary for further processing. Results showed that Support Vector Machine was the first preferred Marijuana/Cannabis drug leaf classification, whereas end-to-end Convolutional Neural Network was second preferred Marijuana/Cannabis drug leaf classification using Artificial Intelligence. Through the series of experimental factor analysis, the newly developed scale was revised to include two types of image detection and classification investigations. Experimental implications of the results are discussed. The proposed marijuana drug detection and classification technique using machine learning approach identify successfully up to CNN classifier 88% accuracy and using SVM classifier 92% accuracy respectively. Keywords:- Image Detection, Image Processing, Image classification, CNN, SVM	en_US
dc.language.iso	en_US	en_US
dc.subject	INFORMATION TECHNOLOGY	en_US
dc.title	MARIJUANA DRUG DETECTION AND CLASSIFICATION USING MACHINE LEARNING APPROACH	en_US
dc.type	Thesis	en_US