För- och nackdelar med bildbehandling för analys av effekten av en trafiksäkerhetsåtgärd

The fact that there are no crashes during every three-year period on every spot of a road network does not mean that the long-term safety level is exactly zero crashes. However, what the lack of recorded crashes does mean is that we need less blunt instruments than crash data for evaluating, for example, the effectiveness of a reconstruction. A new tool, image processing, is presented. It includes reliability tests of an automatic feature for speed measurements and of more detailed semi-automatic features to study the interaction between vulnerable road users and motor vehicle drivers. A case studie from Sweden and Germany is also presented.