Sistem Pendeteksi Kecepatan Kendaraan dengan Menggunakan Metode Deep Learning

Abstract

Object detection based on digital image processing in vehicles is important to be applied in building surveillance systems or as an alternative method of collecting statistical data for effective technical decision making in traffic engineering. This review describes the implementation of vehicle speed detection and measurement using OpenCV. This method involves the use of a cascade classifier to detect cars and a dlib correlation tracker to track objects in traffic video footage. Car detection is performed on each image using a Cascade Classifier, while object tracking uses a correlation tracker to track the location and identity of the vehicle. Each time the position changes between images, the program calculates the car's speed based on the difference in position. Speed ​​is measured in kilometers per hour and is displayed above each passing vehicle. Experimental results demonstrate the program's success in effectively detecting and tracking vehicles, providing clear and accurate visualization of vehicle movement and speed. The program is also capable of measuring performance in frames per second (fps). The conclusions of this journal highlight the potential for further development, optimization, and maintenance as measures to improve system performance in various contexts of use. This journal contributes to the literature in the field of automated traffic monitoring and can be a useful reference for developers of security monitoring and traffic management systems.