摄像头小车寻迹程序，一幅图像采集5行，每行采集80点-Camera car tracing procedures, an image capture five lines, each line collected 80 points

采用tensorflow的深度学习框架，采用opencv，采用时间戳，和特征抓取，自动抓取图片特征，然后存储(Using tensorflow's deep learning framework, opencv, timestamp, and feature capture are used to automatically capture image features, and then store them)