Implementasi Artificial Intelligence Untuk Mendeteksi Rambu Pembatas Kecepatan Rendah Dengan Metode Convolutional Neural Network

Erwin, Yuliansyah (2021) Implementasi Artificial Intelligence Untuk Mendeteksi Rambu Pembatas Kecepatan Rendah Dengan Metode Convolutional Neural Network. Undergraduate Thesis thesis, Institut Teknologi Telkom Purwokerto.

	Text Cover.pdf Download (407kB)
	Text Abstract.pdf Download (7kB)
	Text Abstrak.pdf Download (7kB)
	Text BAB I.pdf Download (145kB)
	Text BAB II.pdf Download (349kB)
	Text BAB III.pdf Download (279kB)
	Text BAB IV.pdf Restricted to Registered users only Download (371kB) | Request a copy
	Text BAB V.pdf Download (7kB)
	Text Daftar Pustaka.pdf Download (140kB)
	Text Lampiran.pdf Restricted to Registered users only Download (407kB) | Request a copy

Abstract

Traffic signs are one of the road equipment that has various forms in the form of pictures, letters, numbers that are used to regulate road users. The population in urban areas, which is estimated to continue to increase to reach 66.6% of the total population in Indonesia, results in a fairly high accident rate due to a lack of public awareness of the applicable traffic sign regulations. This research on the classification of speed limit signs was made to help the public and drivers, especially in urban areas, to recognize speed limit signs. The method used in this study is a convolutional neural network, which is an algorithm in deep learning that imitates the workings of the human brain's nerves to connect with each other and learn patterns. The processes in this research are data collection, preprocessing, and testing the predicted class. The data used is 1064 data which is divided into 3 classes, namely 20km/h, 30km/h, 50km/h. At the preprocessing stage, data is divided, namely 80% training, 10% validation and 10% testing. The literacy process (learning) used is 10, 15, 20 epochs, the highest accuracy results are at epoch 20 it reaches 99% and the lowest accuracy is obtained at epoch 10, which is 94%. Of the 106 data tested, the results from the predicted class show that the test data can be predicted according to its class with the highest accuracy, namely in class 20, which is 97%, the results of this predicted class show that the model can work with an average accuracy of 91.46%. Keywords :Speed sign, classification, convolutional neural network, deep learning, epoch

Actions (login required)

View Item