Video accessing and processing for automatic vehicle guidance

In this paper, an MCU (micro controller unit) oriented automatic guidance system is proposed. Capture strategies including capture analysis and camera setting position determination are discussed. Relationship between setting position and sight view is illustrated for Freescale smart car. Instead of conventional software binarization, a hardware based binary method is proposed to deal with all scan line data for each sample picture in real-time. Keystone distortion correcting algorithm for MCU application improves the performance of track type recognition. Together with rollover preventing calculation, the proposed design is successfully applied to a 1:10 scale model car.