While the interests in intelligent vehicle increase, many people are having concerns about systems that offer the information of distance and relative speed between two cars. This paper presents an algorithm to obtain efficiently the distance between two cars.
IntorductionDriver inattention and poor judgment are the major causes of motor vehicle accidents. Extensive research has shown that intelligent driver assistance systems can significantly reduce the number and severity of these accidents [1]. Compared with vehicles equipped with no crash warning system, vehicles equipped with crash warning systems can reduce more accidents. If the alarm is given a few seconds in advance of the collision, many accidents can be prevented [2]. Therefore, while the interests in intelligent vehicles increase, many people have concerns about systems that offer the information of distance and relative speed between two cars.One of the most popular solutions today is based on radar technologies [3], [4]. The radar technology measures reflect-ions from metal objects, and takes into account the Doppler effect in order to provide relative speed information. However, the high cost of radar systems limits their usefulness.Given that the driving environment is designed around the human driver's ability for visual perception, it may look natural to search for vision solutions. Therefore, another family of solutions is based on image system with two cameras [5], [6]. Such systems are still rather expensive and require accurate calibration between the cameras.The systems based on vision processing with only one camera are recently developed [7-10] and such systems are called "monocular vision system." This paper presents an automobile advance warning algorithm that is carried out using a monocular vision system.Domain limitation via lane identification is proposed in order to efficiently detect advance vehicles. Dark pixel density is used to detect the existence of cars. The distance is calculated using the width between the left and right lanes and the relative speed between two cars is obtained with calculated distance. Finally, the alarm is given a few seconds in advance of the possible collision for drivers to keep safety distance.The proposed algorithm was tested with actual images taken on the road. The 256-gray-level image is used for the simulation. The simulation results show that the maximum error is less than 1 m within the range of 50 m and less than 4 m within the range of 70 m. The simulation results show that more accurate results can be obtained by the proposed algorithm compared with the existing algorithms.The proposed algorithm was implemented using a SoC kit. The included processor is the ALTERA Excalibur ARM EPXAnF1020C3. The image and the calculated distances are displayed via TFT-LCD and a 7-segment, respectively. Figure 1 shows the proposed algorithm for automobile advance alarm systems. Fig.1. Algorithm for automobile advance warning system. This work was supported by the second stage of Brain Korea 21 Project 2007 ...