Image Restoration and Helmet Usage Violation Detection using the Lucy-Richardson Deconvolution Algorithm and Convolutional Neural Network (CNN) Method

M.Redo Anugroho; Lastri Widya Astuti; Indah  Permatasari

doi:10.59934/jaiea.v5i1.1732

Authors

M.Redo Anugroho Universitas Indo Global Mandiri
Lastri Widya Astuti Universitas Indo Global Mandiri
Indah Permatasari Universitas Indo Global Mandiri

DOI:

https://doi.org/10.59934/jaiea.v5i1.1732

Keywords:

Restorasi Citra, Lucy-Richardson Deconvolution, CNN, Deteksi Helm, Klasifikasi Citra, ETLE

Abstract

The high number of traffic violators on the road due to motorcyclists not wearing helmets remains a serious problem, despite the implementation of the electronic ticketing system (ETLE). One of the challenges in detecting these violations is the poor image quality caused by blurring. This study proposes a combined method using the Lucy-Richardson Deconvolution restoration algorithm and classification using a Convolutional Neural Network (CNN) to improve the accuracy of violation detection. The dataset used consists of 1,101 images extracted from videos, divided into two classes: wearing a helmet and not wearing a helmet. The images were tested with and without the restoration process to compare detection performance. After undergoing restoration and CNN model training, the system achieved an accuracy of 93% on the test data. The detection results showed a significant improvement compared to images without restoration, where accuracy only reached 87%. In addition, the system can process blurred images and classify helmet objects more accurately based on evaluation using a confusion matrix and bounding box visualisation. Thus, the integration of the Lucy-Richardson Deconvolution algorithm with CNN has proven to be effective in improving image quality and helmet violation detection accuracy, and has the potential to be applied to ETLE systems to support automated traffic law enforcement.

Downloads

Download data is not yet available.

References

M. Dwi, O. S. Matompo, and I. Lestiawati, “Penerapan Sanksi Denda Terhadap Pengendara Motor Yang Tidak Menggunakan Helm (Studi Pada Kepolisian Resort Sigi),” J. Sasi, vol. 16, no. 22, pp. 1903–1917, 2019.

SIMATA, “Satu Data Provinsi Sumatera Selatan,” Sist. Inf. Satu Data Sumsel, no. 1, p. 34, 2021.

S. Aprita and M. Nurlani, “Penegakan Hukum Lalu Lintas Melalui E-Tilang Dalam Meningkatkan Kesadaran Hukum Berlalu Lintas Di Kota Lubuk Linggau,” Karya J. Pengabdi. Masy., vol. 1, no. 2, pp. 54–64, 2024, doi: 10.70656/jpkm.v1i2.174.

M. Cahyanti, M. Ridwan, and D. Septian, “A Real-Time Helmet Detection System Based on YOLOv8 to Support Traffic Law Enforcement Sistem Deteksi Penggunaan Helm Secara Real-Time Berbasis YOLOv8 untuk Mendukung Penegakan Hukum Lalu Lintas,” vol. 29, no. 1, pp. 1–10, 2025, doi: 10.46984/sebatik.v29i1.2585.

R. Gustriansyah, J. Alie, A. Sanmorino, R. Heriansyah, and M. N. Megat Mohamed Noor, “Machine Learning for Clustering Regencies-Cities Based on Inflation and Poverty Rates in Indonesia,” Indones. J. Inf. Syst., vol. 5, no. 1, pp. 64–73, 2022, doi: 10.24002/ijis.v5i1.5682.

Z. R. Mair and M. A. Rahmanda, “Perbandingan Versi Terbaik YOLO Dalam Mendeteksi Jarak Spasi Antar Baris Tulisan Tangan,” J. Sains, Nalar, dan Apl. Teknol. Inf., vol. 4, no. 2, pp. 103–110, 2025, doi: 10.20885/snati.v4.i2.40414.

K. Khoiriyah and M. F. A. A. Armawan, “Deteksi Pengendara Motor Tanpa Menggunakan Helm Dengan Algoritma Deep Learning Yolo,” Jeis J. Elektro Dan Inform. Swadharma, vol. 3, no. 2, pp. 72–82, 2023, doi: 10.56486/jeis.vol3no2.360.

S. B. Setyawan, W. Pribadi, H. Arrosida, and E. P. Nugroho, “Sistem Deteksi Pengendara Sepeda Motor Tanpa Helm dan Kelebihan Penumpang pada Dengan Menggunakan YOLO V3,” Semin. Nas. Terap. Ris. Inov., vol. 7, no. 1, pp. 430–438, 2021.

Albert, K. Gunadi, and E. Setyati, “Deteksi Helm pada Pengguna Sepeda Motor dengan Metode Convolutional Neural Network,” J. Infra, vol. 8, no. 1, pp. 295–301, 2020.

M. Harahap, J. P. Sihombing, and A. R. Lubis, “Deblurring Citra Dengan Metode Lucy Richardson Deconvolution,” vol. 3, no. April, pp. 441–454, 2020.

Nurohman, Rudi Heriansyah, Dwi Asa Verano, and Zaid Romegar Mair, “Deteksi Penyakit Diabetes Retinopathy Menggunakan Citra Digital Dengan Metode Convolutional Neural Network (Cnn),” Pros. Snast, no. November, pp. 311–320, 2024, doi: 10.34151/prosidingsnast.v1i1.5120.

“Kustoyo_2018110072.pdf-1.pdf.”

Pratama M. Duta, Gustriansyah Rendra, and Purnamasari Evi, “Klasifikasi Penyakit Daun Pisang Menggunakan Convolutional Neural Network (CNN),” J. Teknol. Terpadu, vol. 10, no. 1, pp. 1–6, 2024.