Real-time Image Identification for Vehicle Positioning System

An innovative vehicle real-time positioning method was developed to maximize the use of existing visible light camera systems on vehicles, especially in areas without Global Positioning System (GPS) signals or blind spots. This approach eliminates the need for additional positioning devices and leverages the potential of existing vehicle driver assistance systems. As vehicles are equipped with cameras for monitoring the surrounding environment, we optimized the use of these existing resources for an efficient and accurate innovation in vehicle positioning. Firstly, we used cameras in four directions (the front, rear, left, and right) of the vehicle to capture images. By integrating images in four directions and using a transformation algorithm, a complete top-down view was formed with geographic information to enhance positioning accuracy. The integrated top-down view image is useful for obtaining information on surrounding environments and identifying the vehicle's real-time location in parking lots. The developed method serves as a foundation for future applications in automated driving planning and positioning.