Over the last three decades, the global population of human beings has increased at an exponential rate, resulting in an equal rise in the number of vehicles owned and used globally. Vehicle traffic is a major economic component in both urban and rural areas, and it requires proper management and monitoring to ensure that this mass of vehicles coexists as smoothly as possible. The amount of vehicular traffic on roads around the world, with Nigeria as a case study, results in varying degrees of traffic rule violations, especially red light jumping.  To arrest offenders and resolve the weaknesses and failures of human traffic operators who cannot be everywhere at once, efficient traffic violation and number plate recognition systems are needed. There are several methods for reading characters, which can be alphabets, numbers, or alphanumeric. To minimize processing time and computational load on the machine, this research proposed k-Nearest Neighbour for plate number character recognition. The system was developed and evaluated. From the result, the localization of license plate regions within an image was 92 percent accurate, and character recognition was 73 percent accurate.