Simultaneous Traffic Sign Recognition and Real-Time Communication using Dual Camera in ITS

Research on intelligent transportation system (ITS) is increasing owing to its incredible potentiality in transportation. Numerous features are added in the in-road vehicles facilitating the utilization of newly-developed technologies to reduce traffic collisions and assure human safety. Among them, camera-mounted smart cars are currently very common. These cameras can be used to receive data from light-emitting diodes (LED) of other vehicles, traffic signals or roadside units, which is termed as optical camera communication (OCC). Another significant task that can be performed using the cameras is the automatic recognition of traffic signs. Deep learning algorithms are comprehensively developed for the detection of the LEDs or signs. However, both communication and recognition at the same time is a challenging task as it requires complex image-processing techniques to process the LED and sign images simultaneously. Motivated by this problem, we propose a dual-camera system and an algorithm for communication and recognition at the same time without modifying the current transportation system. Convolutional neural network is used to detect the desired objects primarily. Then one of the cameras is assigned to capture the image frames for further processing of the communication or recognition mechanism. Our algorithm will ensure the reduction of overall computational complexity. At the end of the paper, we enlist the challenges that should be envisaged while considering our algorithm.