<![CDATA[In Indonesia, the dominant use of transportation at present is road-mode land transportation compared to other transportation namely sea and air transportation. Accident rates were found to be increasing, due to lack of awareness of driving safety and security. The dominant human error factor is the cause of an accident. One of the accident factors is caused by drivers who lose control, because they ignore the maximum and minimum speed limiting signs. The solution to this problem is to create a system that can recognize maximum and minimum speed limiting signs. The research applies the Histogram of Oriented Gradients (HOG) method to obtain the characteristic feature extraction from the signs, then classifies the signs using the K-Nearest Neighbor (KNN) method. The system requires a raspberry pi camera to capture images for detection and object recognition. If the system manages to recognize the signs according to the actual conditions traversed by the driver, it will get notification of speed sign figures in the form of sound from the speakers. System testing is done based on a varied distance that is a distance of 3m, 5m, 7m, and 9m. The four distances that are the best distances in detecting speed limiting signs are 5 meters. The average result of recognition / recognition recognition accuracy using HOG method based on the best detection distance is 97.91%. Classification testing using K-NN consists of 650 training data and 48 test data obtained k = 1 and k = 2 accuracy values ​​of 97.91%, accuracy of k = 3, k = 4, and k = 5 values ​​of 95.83 %. The average time of computing the system to recognize objects 897 milliseconds.]]>
Copyrights © 2020
