Deteksi Posisi Plat Nomor Kendaraan Menggunakan Metode Transformasi Hough dan Hit or Miss

Yuda Puspito (Department of Electrical Engineering Faculty of Engineering University of Lampung Jl. Sumantri Brojonegoro 1, Bandar Lampung)
FX Arinto Setyawan
Helmy Fitriawan

Electrician : Jurnal Rekayasa dan Teknologi Elektro
Vol. 12 No. 3 (2018)

Abstrak: Penelitian ini dikembangkan sebuah sistem pendeteksi posisi plat nomor kendaraan yang ditampilkan pada GUI Matlab. Pendeteksian posisi plat nomor kendaraan menggunakan dua metode, yaitu metode transformasi hough dan transformasi hit or miss. Tahap pengolahan citra yang digunakan meliputi: binerisasi, aras keabuan, deteksi tepi, pemotongan citra, filtering, dan resizing. Keefektifan sistem ini diukur dengan perhitungan terhadap nilai perolehan (recall) dan nilai ketepatan (precision). Berdasarkan hasil penelitian didapatkan bahwa sistem berhasil mendeteksi posisi plat nomor kendaraan dengan tingkat keberhasilan pendeteksian sebesar 76% untuk nilai threshold 0,75, 72% untuk nilai threshold 0,8 dan 48% untuk nilai threshold 0,85. Hasil penelitian juga menunjukkan nilai rata-rata recall sebesar 54% untuk nilai threshold 0,75, 50% untuk nilai threshold 0,8 dan 40% untuk nilai threshold 0,85, sedangkan nilai rata-rata precision sebesar 14% untuk nilai threshold 0,75, 14% untuk nilai threshold 0,8 dan 12% untuk nilai threshold 0,85.Kata kunci: Trasnformasi Hough, Transformasi Hit Or Miss, Recall, Precission. Abstract: This research was developed a detection system of vehicle license plates that displayed at the GUI Matlab. The detecting the number plate position of the vehicle uses two methods, namely the transformation method of hough and the transformation of hit or miss. The image processing stages used include: binerization, gray level, edge detection, image cutting, filtering, and resizing. The effectiveness of this system is measured by calculating the value of the recall and the precision. Based on the results of the study it was found that the system successfully detected the number plate position of the vehicle with a detection rate of 76% for the threshold value of 0.75, 72% for the threshold value of 0.8 and 48% for the threshold value of 0.85. The results also showed an average recall value of 54% for the threshold value of 0.75, 50% for the threshold value of 0.8 and 40% for the threshold value of 0.85, while the average value of precision was 14% for the threshold value of 0.75, 14% for the threshold value of 0.85, while the average value of precision was 14% for the threshold value of 0.75, 14% for the threshold value of 0.8 and 12% for the threshold value of 0.85.Keywords: Trasnformasi Hough, Transformasi Hit Or Miss, Recall, Precission.