Intelligent traffic cone based on vehicle accident detection and identification using image compression analysis and RFID system
Songkran Kantawong, Tanasak Phanprasit · International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology · 2010
The proposed intelligent RFID traffic cone for vehicle accident detection and identification avoids problems that usually arise with normally vehicle crash reporting systems, especially those related to image processing and insurance techniques. This RFID technique deals with a multi-vehicles, multi lane and multi road even or junction area. It provides an efficiency time management scheme with correct data reporting, in which a dynamic time schedule is worked out in real time for the driver or passengers of each accident situations. The time operation of the system emulates the judgment of a traffic policeman on duty or user that may have PDA nearby RFID traffic cone. The image compression present here is used along with RFID information to get a precise event data that composed of image encoding and decoding algorithms called wavelet transform with principle component analysis (PCA) via vector quantization techniques (VQ). The small bit rates for high-speed data transmission with a small space for data storage are required on wireless transmission channel. Simultaneously, the peak signal to noise ratio (PSNR) has to be maintained. The traffic management system model is constructed for testing on traffic lights, vehicles transit and traffic cone with RFID solution system. By applying the proposed technique, performance has been improved which indicated by lower bit rate and better PSNR for image compression algorithm. The RFID data records of vehicles can be sent to the traffic information center on wide area network.