Implementasi Algoritma Convolutional Neural Network dengan Arsitektur Mobilenetv2 untuk deteksi limbah bahan berbahaya dan beracun (B3)

Afrinaldi, Afrinaldi (2023) Implementasi Algoritma Convolutional Neural Network dengan Arsitektur Mobilenetv2 untuk deteksi limbah bahan berbahaya dan beracun (B3). Sarjana thesis, UIN Sunan Gunung Djati Bandung.

[img]
Preview
Text (COVER)
1_cover.pdf

Download (112kB) | Preview
[img]
Preview
Text (ABSTRAK)
2_abstrak.pdf

Download (145kB) | Preview
[img]
Preview
Text (DAFTAR ISI)
3_daftarisi.pdf

Download (108kB) | Preview
[img]
Preview
Text (BAB I)
4_bab1.pdf

Download (149kB) | Preview
[img] Text (BAB II)
5_bab2.pdf
Restricted to Registered users only

Download (401kB) | Request a copy
[img] Text (BAB III)
6_bab3.pdf
Restricted to Registered users only

Download (663kB) | Request a copy
[img] Text (BAB IV)
7_bab4.pdf
Restricted to Registered users only

Download (1MB) | Request a copy
[img] Text (BAB V)
8_bab5.pdf
Restricted to Registered users only

Download (77kB) | Request a copy
[img] Text (DAFTAR PUSTAKA)
9_daftarpustaka.pdf
Restricted to Registered users only

Download (162kB) | Request a copy

Abstract

INDONESIA : Limbah merupakan bahan yang tidak terhindarkan dalam kehidupan manusia sehari-hari dan meningkat seiring dengan pertumbuhan populasi serta kegiatan manusia. Terdapat beberapa jenis limbah yang umum dikenal, yaitu limbah anorganik, organik, dan limbah Bahan Berbahaya dan Beracun (B3). Limbah Bahan Berbahaya dan Beracun (B3) memiliki dampak serius terhadap lingkungan dan kesehatan manusia karena mengandung zat kimia berbahaya. Oleh karena itu, penting sekali untuk mengelola limbah dengan benar guna mencegah dampak buruk yang dapat terjadi. Salah satu langkah awal dalam pengelolaan limbah adalah identifikasi yang tepat terhadap jenis limbah Bahan Berbahaya dan Beracun (B3). Identifikasi ini dapat dilakukan dengan menggunakan teknologi image recognition yang memanfaatkan kemampuan komputer untuk mempelajari fitur-fitur dari gambar. Dalam penelitian ini, dilakukan klasifikasi limbah Bahan Berbahaya dan Beracun (B3) menggunakan algoritma Convolutional Neural Network (CNN) dengan arsitektur MobileNetV2. Dataset yang digunakan terdiri dari 3940 gambar yang terbagi dalam 7 kelas diantaranya carcinogenic tetragenic mutagenic, dangerous for environment, harmful, infectious, pressure gas, toxic, dan non hazard. Sebelum melatih model, dilakukan pra-pemrosesan data seperti pelabelan data, augmentasi data, pemisahan data, dan normalisasi data. Dengan menggunakan arsitektur MobileNetV2, model yang dikembangkan menghasilkan hasil yang baik berdasarkan perhitungan confussion matrix dimana accuracy yang didapatkan sebesar 88%, precision 90%, recall 88%, dan f1-score 88%. Hasil ini didapatkan setelah menguji model dengan 20 epoch dan unfreeze pada base model. ENGLISH : Waste is an inevitable material in human daily life and increases along with population growth and human activities. There are several types of waste that are commonly known, namely inorganic waste, organic waste, and hazardous and toxic waste. Hazardous and toxic waste has serious impacts on the environment and human health because it contains harmful chemicals. Therefore, it is very important to manage waste properly to prevent the adverse effects that may occur. One of the initial steps in waste management is the proper identification of the type of hazardous and toxic waste. This identification can be done by using image recognition technology that utilizes the computer’s ability to learn the features of images. In this study, hazardous and toxic waste classification was performed using the Convolutional Neural Network (CNN) algorithm with the MobileNetV2 architecture. The dataset used consists of 3940 images divided into 7 classes including carcinogenic tetragenic mutagenic, dangerous for environment, harmful, infectious, pressure gas, toxic, and non hazard. Before training the model, data preprocessing was performed such as data labeling, data augmentation, data splitting, and data normalization. By using the MobileNetV2 architecture, the developed model produced good results based on the confussion matrix calculation where the accuracy obtained was 88%, precision 90%, recall 88%, and f1-score 88%. These results were obtained after testing the model with 20 epochs and unfreezing on the base model.

Item Type: Thesis (Sarjana)
Uncontrolled Keywords: Convolutional Neural Network (CNN); Confussion matrix; Klasifikasi; Limbah Bahan Berbahaya dan Beracun (B3); MobileNetV2; Pra-premrosesan data;Classification;Data preprocessing;Hazardous and toxic waste
Subjects: Systems > Computer Modeling and Simulation
Technology, Applied Sciences
Divisions: Fakultas Sains dan Teknologi > Program Studi Teknik Informatika
Depositing User: Afrinaldi Afrinaldi
Date Deposited: 08 Sep 2023 03:42
Last Modified: 11 Sep 2023 00:35
URI: https://digilib.uinsgd.ac.id/id/eprint/76176

Actions (login required)

View Item View Item