20.500.12556/RUL-87763 Zaznava prevožene rdeče luči na podlagi algoritmov računalniškega vida Detecting Red Light Runs Using Computer Vision Algorithms Raziskave s področja cestne varnosti so pokazale, da je mogoče število prekrškov vožnje skozi rdečo luč in posledično prometne nesreče, katerim botruje tovrsten prekršek, znatno zmanjšati, s postavitvijo sistemov za samostojno zaznavo prekrškarjev. Osrednja tema tega dela je prometen nadzorni sistem na osnovi računalniškega vida, ki zazna spremembo semaforja brez signala semaforskega krmilnika ter zazna prevoženo rdečo luč. Ob takšnem dogodku sistem samodejno posname in shrani prekršek. Poudarek je na kompaktnosti in preprosti postavitvi sistema, saj je to nadgradnja obstoječih rešitev, ki so se izkazale za nepraktične ravno zaradi kompleksnosti vgradnje. Sistemska rešitev je pripravljena z nadzorno kamero Axis P1357 in mikroračunalnikom Raspberry Pi 2, v programskem jeziku Python z uporabo knjižnice OpenCV. V okviru naloge sem opisali izzive, s katerimi sem se soočal pri programiranju in rešitve, ki sem jih pripravil, ter podal mnenje, kaj in kako bi se dalo izboljšati. Vodilo naloge je bilo pripraviti rešitev za nadzor prometa z opremo, ki je bila dostopna v laboratoriju. Researchers have shown, that the rate of traffic lights violations and thus related accidents can be significantly reduced by introducing autonomous red-light cameras to crossroads. The main topic of this work, is autonomous traffic control system for intersections, based on computer vision, that can detect change in traffic lights, without any signals from traffic lights controller and can identify a red light runner. That triggers the system to record and save the footage of the event for later inspection. Emphases of the project is to create a compact, easy to install system, that upgrades on the now existing systems, that need modifications of road surfaces and traffic lights controllers. Solution is made with a common security camera AXIS P1357 and microcomputer Raspberry Pi 2, in Python whit the use of OpenCV library. As a part of this work I discuss challenges and solutions, to the problems I encountered and I give a few suggestions as to what could be done better. The main thread of this work was to create a solution for traffic control with the equipment found in the laboratory. cestna varnost prometni nadzor računalniški vid semafor road safety traffic control computer vision traffic lights true false false Slovenski jezik Angleški jezik Magistrsko delo/naloga 2016-12-13 07:44:36 2016-12-13 07:44:40 2022-08-09 12:25:11 0000-00-00 00:00:00 2016 0 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 1970-01-01 33661 Susteric_Miha_-_Zaznava_prevozene_rdece_luci_na_podlagi_algoritmov_racunalniskega_vida.pdf Susteric_Miha_-_Zaznava_prevozene_rdece_luci_na_podlagi_algoritmov_racunalniskega_vida.pdf 1 DF89480C70B0A78EABAF7DE16C8898F6 b9e3b7c81f3ef0dd482aa6f377320163d806922bbc9d6b47010992fd86442d79 1ce378c6-a1b1-11eb-a523-00155dcfd717 20.500.12556/rul/0a0f372d-eca2-4f4a-8ae2-524dba19b369 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=91284 Fakulteta za elektrotehniko 0 0 0