Collision risks detection and intelligent speed control

U ovom diplomskim radom razvijen je algoritam koji na temelju udaljenosti vozila koje se nalazi ispred ili iza vozila na kojemu je implementiran algoritam, računa brzinu te poduzima potrebnu radnju. U predloženom algoritmu korišten je Viola-Jones detektor koji služi za detekciju vozila nakon čega se proračunava udaljenost vozila. Viola-Jones detektor je istreniran na ranije prikupljenim slikama nakon čega je testiran na novom skupu slika. Iz rezultata testiranja može se vidjeti kako algoritam radi prilično dobro (vrlo rijetko se dešavaju netočne detekcije i detekcije kojima je IoU manji od 70%).In this graduate thesis, an algorithm was developed which, based on the distance of the vehicle which stands in front of or behind the vehicles on which the algorithm was implemented, calculate speed and takes some action. In the proposed algorithm, the Viola-Jones detector for vehicle detection is used. The Viola-Jones detector was trained on previously acquired images, after which is tested on a new set of images. The test results show that the algorithm works quite well (very rarely there are inaccurate detections and detections with IoU less than 70%)