EKSTRAKSI DAN PENGENALAN PLAT NOMOR KENDARAAN BERMOTOR DI INDONESIA

AVIANTO, DONNY (2014) EKSTRAKSI DAN PENGENALAN PLAT NOMOR KENDARAAN BERMOTOR DI INDONESIA. S2 thesis, UAJY.

	Text (Halaman Judul) MTF002024.pdf Download (753kB)
	Text (Bab I) MTF102024.pdf Download (65kB)
	Text (Bab II) MTF202024.pdf Download (276kB)
	Text (Bab III) MTF302024.pdf Restricted to Registered users only Download (571kB)
	Text (Bab IV) MTF402024.pdf Restricted to Registered users only Download (1MB)
	Text (Bab V) MTF502024.pdf Download (785kB)

Abstract

License plate recognition is a technology that uses image processing techniques to automatically identify vehicles from the license plate image. This technology can be applied in various applications such as tracking stolen cars, automatic parking ticket payments, and traffic statistics collection. In this research, will be discussed a method to recognize Indonesian vehicles license plates, which consists of 3 main parts: license plate extraction, character segmentation, and character recognition. In the first part, used morphological opening operation to generate candidate image plate, and vertical and horizontal projection method to obtain the actual location of number plate. In the second part, the system will segmenting each character of license plate image from extraction result using the scan-lines. In the last section, each segmented image will be transformed using Haar Wavelet to reduce the dimension and recognized by Momentum Backpropagation Neural Network algorithm. To conduct a trial, 35 images are used as test images with various lighting condition and background. Based on the results of experiments performed, the system is able to perform extraction and segmentation of license plate characters well on all test images. While on the character recognition, from the 276 character tested, the system is able to recognize 268 character of them so accuracy rate reaches 97.10%.

Actions (login required)

View Item