Solve traffic sign recognition challenges
Article By : Jiji Gangadharan
An unnoticed sign board can cause fatal accidents, but how many times do we miss the traffic signs during a long drive? Is it that difficult to see the traffic sign?
An unnoticed sign board can cause fatal accidents, but how many times do we miss the traffic signs during a long drive? Is it that difficult to see the traffic sign?
There are many reasons why we miss the traffic signs. Wavering concentration, tiredness, sleep deprivation are few common reasons. Some of the other factors that contribute to missing the signals are poor illumination, an influence of the exterior environment and weather conditions. Yet it seems illogical to have a person next to us warning us about the traffic signs. It is a lot more acceptable to instead have a system that can detect the traffic signs to guide and warn the driver.
Camera-based traffic-sign recognition systems identify signs in real-time by processing the videos that are captured through a front-facing camera mounted on a vehicle. They help the driver by providing warnings, commands and sometimes by taking control of the vehicle itself. For example, in a 60 km/hr speed zone, if the driver exceeds that speed, the system gives warnings. The system can coordinate with the speed governor and other circuits to reduce the speed if the driver continues to drive at, say 80 km/hr, even after an 'X' number of warnings.
![[traffic-sign-recognition figure 1]](/wp-content/uploads/sites/4/2020/04/traffic-sign-recognition_figure1_optimised.jpg)
Figure 1: The traffic sign recognition system concept.
A vision-based traffic sign recognition system has mainly two modules: the detection module and the recognition module. The detection module localises the sign region from the image/video, and the recognition module identifies what the sign is. In the detection phase, the highly probable sign regions are extracted and given to the recognition system to identify the sign.
Traffic signs have country specific standards with defined shapes and specific combination of colours. The most preferred are "pure" colours for better visibility. Most important feature considering vision algorithm is their static property. Despite all these positive factors, there are 'N' number of factors that make the traffic sign recognition challenging.