双目深度估计

基本思路：
１．双目摄像头标定
２．获取视差图
３．获取三维深度
４．同过三维深度信息再去做其他（如活体）
https://blog.csdn.net/xiao__run/article/details/78887362
https://blog.csdn.net/chentravelling/article/details/70254682
https://blog.csdn.net/Taily_Duan/article/details/52165458
https://blog.csdn.net/chentravelling/article/details/53671279
https://blog.csdn.net/u010369450/article/details/78787805
https://www.cnblogs.com/tjulxh/archive/2012/06/14/2548977.html

https://blog.csdn.net/u014283958/article/details/52673304

https://blog.csdn.net/woniu199166/article/details/79423138