TRAFFIC RULES VIOLATION DETECTION SYSTEM BY USING DEEP LEARNING

Abstract

Real time identification systems are very important and needful for safety, security rule following and socialism and also for own safety concerns. Traffic rules are important for safety as traffic laws are to prevent drivers of vehicles from causing accidents or hitting pedestrians. They are also to help control the flow of traffic so that it is more efficient. Traffic Rule Violations are leading cause of accidents, according to WHO India is a leading country in casualties occurring on road. The current system uses human interaction for rule violation detection, as it is a manual process it has some limitations, on multiple occasions we find the system gets corrupt. An alternative solution would be AI-developed System. With our system, we can detect multiple rule violations, for example, Vehicle crossing signal during red light or driving without a helmet, etc. Basic idea is to detect these violations through preinstalled cameras. We can do it by ML based algorithm where we can detect the violators by ImageProcessing, getting the number pate, categorizing violation accordingly and issuing fine. Which will help increase the efficiency of traffic rule enforcement