Deteksi Helm Keselamatan Kerja Menggunakan CNN

ulwan, muhammad rasyid (2022) Deteksi Helm Keselamatan Kerja Menggunakan CNN. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 07211840000073-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (22MB)

Abstract

Helm keselamatan kerja merupakan salah satu APD yang berfungsi untuk melindungi kepala dari segala bentuk hantaman langsung ke kepala penggunanya. Terdapat aturan pemerintah yang mengatur dan mewajibkan penggunaan Alat Pelindung Diri dimana salah satunya yaitu Helm Keselamatan Kerja seperti pada Peraturan Menteri Tenga Kerja dan Transmigrasi Republik Indonesia NOMOR PER.08/MEN/VII/2010 tentang ALAT PELINDUNG DIRI. Tetapi hal itu tidak menjamin semua pekerja mengenakan Helm Keselamatan Kerja walaupun sudah ada instruksi untuk digunakan. Masih sering didapati pekerja yang mengabaikan keselamatan kerja. Perusahaan - perusahaan yang melakukan pekerjaan pada umumnya sudah mengerahkan pengawas yang bertugas untuk mengawasi penggunaan APD dimana pengerahannnya juga sudah ada aturan yang mengatur. Tetapi pengawas masih melakukan pengawasan secara manual yang dimana memiliki limitasinya sendiri seperti luas area yang dapat diawasi dan banyaknya jumlah pekerja yang harus diawasi. Maka dari itu, dalam penelitian ini diambil suatu tujuan yaitu Merancang system yang dapat mendeteksi penggunaan helm keselamatan kerja secara real-time. Setelah dilakukan pengujian, didapatkan hasil pengujian model dengan nilai precision 0,92, recall 0,87, mAP@.5 0,9 dan untuk pengujian sistem trigger alarm didapatkan akurasi palinG rendah 0,82.
==============================================================================================================================
Safety Helmet or Hardhat is one of the Personal Protective Equipment which purpose is to protect the wearer from direct impact to the head. There are regulations that stated the importance and obligated the use of Personal Protective Equipment in which Safety Helmet is included such as Regulation of Republic of Indonesia Ministry of Manpower NOMOR PER.08/MEN/VII/2010 for PERSONAL PROTECTIVE EQUIPMENT REGULATION. Despite the obligation of the use of safety helmet as one of the Personal Protective Equipment, it does not guarantee all the field workers will wear the helmet. Companies that held the constructions usually had already deployed supervisor to ensure that all worker wear the safety helmet. The deployment of the supervisor itself is also regulated in one of the regulations from Republic of Indonesia Ministry of Manpower. But the method that is used to do supervision is still done manually by the supervisor which has its limitations. The vast area of the construction sites and the number of personnel that is more than a human can count is a challenge for a human supervisor to carry on their duty to supervise every personnel in the area. Therefore, this research aims to develop a system that is capable of automatically detecting the personnel that wears safety helmets and the ones that do not wear a safety helmet and triggering a sort of alarm when the system detects personnel that does not wear a safety helmet properly. The development of the system will be utilizing a Convolutional Neural Network, mainly designed to do 2D recognition. After conducting tests, the results for the performance of the model is 0.92 for precision, 0.87 for recall , 0.92 for the mAP@.5, and 0.82 for the accuracy of the detection system.

Item Type:	Thesis (Other)
Additional Information:	RSKom 006.32 Ulw r-1 2022
Uncontrolled Keywords:	ConvLSTM2D, remote Photoplethysmography (rPPG), Detak Jantung, Machine Learning. ConvLSTM2D, Remote photoplethysmography (rPPG), Heart Beat, Machine Learning.
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science. EDP
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Computer Engineering > 90243-(S1) Undergraduate Thesis
Depositing User:	Mr. Marsudiyana -
Date Deposited:	17 Jun 2026 06:25
Last Modified:	17 Jun 2026 06:25
URI:	http://repository.its.ac.id/id/eprint/133857

Actions (login required)

View Item