Image Processing Application in Toll Collection


  Image Processing Application in Toll Collection.doc (Size: 1,014 KB / Downloads: 1)

INTRODUCTION

The toll rate charged for a vehicle in using a tunnel or a bridge is usually proportional to number of axles of a car. For example, for the Eastern Cross Harbour Tunnel of Hong Kong, each additional axle is charged for $3 USD. Currently, the toll is collected by an operator, who is responsible to determine the axle number when the vehicle arrives at the toll booth. However, the booth operator has a limited angle of view and sometimes the operator may be blocked by the vehicle as well. It is sometimes difficult to determine the correct number of axles. In most cases, operators determine the axle number based on their experiences. It is, therefore, desirable to have an automatic system to either assist the booth operator to count the axle well in advance or to verify that the toll is correctly charged.



THE IMAGE PROCESSING SYSTEM FOR WHEEL RECOGNITION

Initially, the system will be used as a tool to validate the fare charged. The setup cost is a major design criterion to be considered and therefore, components of the system are low-cost commodity products. The basic setup includes a low-end digital video camera and a computer. The resolution of the image produced by the camera is only 320x240 pixels. Since it is difficult to examine axles of a vehicle and therefore, the system will detect wheels instead of axles. The camera is located so that wheels of a vehicle can be captured when it passes through the toll booth. Figure 1 shows the position of the camera relative to the road and a sample image captured by the camera is given in Figure 2. From observations, it takes about two seconds for the front wheel and the rear wheel of a car to pass through the camera, implying that the processing system must complete the recognition process within the two-second time frame.