Recognition method of license plate for black box video using Apache Kafka

Abstract

With the increase in the number of images from various channels such as a vehicle's black box, CCTV and drone, the amount of data related to the vehicles and the traffic situations is also explosively increasing. Also, the greatly increased image data analysis is further emphasizing the importance of large-scale image data processing. In this regard, this study proposed and designed a distribution system using Kafka. For high-speed processing of data, black box image frames were distributed to several nodes using Kafka, a real-time message distribution processing system. Moreover, by setting the frame of black box image as input to each node, the localization of car license plates through image processing, and segmented character recognition were explored.