未验证 提交 e1ae368e 编写于 作者: 朱本福 提交者: GitHub

Update README.md

上级 a1a25308
......@@ -11,7 +11,7 @@ works in real-time with detection and recognition accuracy up to 99.8% for Chine
mtcnn检测到车牌之后,通过回归得到四个角点,做透视变换对齐得到水平车牌,实测可以处理角度非常偏的车牌,
-------
![image](https://github.com/zhubenfu/License-Plate-Detect-Recognition-via-Deep-Neural-Networks-accuracy-up-to-99.9/blob/master/%E5%BE%AE%E4%BF%A1%E5%9B%BE%E7%89%87_20180720093811.png)
单个车牌字符分割是基于haar+cascade加上逻辑筛选,
--------
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册