对图片上的特征物可以实现框选或者是描边提取，同时记录坐标，可用于机器学习。(The feature of the picture can achieve frame selection or stroke extraction, while recording the coordinates, can be used for machine learning.)