Electronic Theses and Dissertation

Dewasa ini, kurangnya kewaspadaan masyarakat dalam tikungan jalan menyebabkan seringnya terjadi kecelakaan. implementasi perangkat deteksi kendaraan dapat membantu untuk mengurangi persentase kecelakaan yang terjadi pada tikungan jalan. beberapa penelitian sebelumnya, sudah melakukan penelitian terkait deteksi kendaraan menggunakan metode blob (binary large object), haar-like feature, serta motion blur image. namun, proses deteksi selama ini masih sering terkendala seperti bentuk objek, kesalahan dalam mendeteksi, serta metode yang digunakan kurang efektif. oleh karena itu, penelitian ini merancang perangkat deteksi kendaraan di tikungan jalan berbasis opencv. opencv ditujukan untuk memproses frame-frame dalam video rekaman cctv sehingga objek kendaraan dapat terdeteksi secara otomatis. hasil penelitian merupakan perancangan perangkat yang dapat mendeteksi kendaraan pada tikungan jalan dan saat kendaraan terdeteksi, maka lampu akan menyala, dengan accuracy 73,15%, precision 90,9% dan recall 75%. kata kunci: kecelakaan, perangkat pendeteksi kendaraan, opencv

Nowadays, the lack of public awareness at road corners causes frequent accidents. Implementing vehicle detection devices can help to reduce the percentage of accidents that occur on road curves. Several previous studies have conducted research related to vehicle detection using the BLOB (Binary Large Object), Haar-like feature, and Motion Blur Image methods. However, the detection process so far is still often hampered by problems such as the shape of the object, errors in detection, and the methods used are less effective. Therefore, this research designs a vehicle detection device at road corners based on OpenCV. OpenCV is intended to process frames in CCTV video recordings so that vehicle objects can be detected automatically. The result of the research is the design of a device that can detect vehicles at road bends and when the vehicle is detected, the lights will turn on, with an accuracy of 73.15%, precision of 90.9% and recall of 75%. Keywords: accident, vehicle detection device, OpenCV