I recently learned opencv and found some source code on the Internet. I found that the effect is not very good. I found that the parameters of the picture will be changed together, otherwise it will not be recognized. So I found HyperLPR on the Internet and now record the environment.
Refer to the blog of the big guy https://blog.csdn.net/sdu_qrt/article/details/120554978?spm=1001.2014.3001.5501
Install the required packages
pip install keras==2.0.9 -i http://pypi.douban.com/simple --trusted-host pypi.douban.com
pip install Theano -i http://pypi.douban.com/simple --trusted-host pypi.douban.com
pip install Scipy -i http://pypi.douban.com/simple --trusted-host pypi.douban.com
pip install opencv-python==3.4.3.18 -i http://pypi.douban.com/simple --trusted-host pypi.douban.com
pip install scikit-image -i http://pypi.douban.com/simple --trusted-host pypi.douban.com
pip install pillow -i http://pypi.douban.com/simple --trusted-host pypi.douban.com
pip install tensorflow==1.2.0 -i http://pypi.douban.com/simple --trusted-host pypi.douban.com
pip install h5py==2.10
pip install numpy==1.16.0 -i http://pypi.douban.com/simple --trusted-host pypi.douban.com
Note that if an error is reported, be sure to check whether the version is consistent with the version I listed here.
HyperLPR installation package
extraction code hin1
Copy the hyperlpr_py3 folder, this folder is based on python3
Copy to D:\Anaconda\anzhuang\envs\HyperLPR\Lib This directory is the virtual environment directory you created and paste it, and change the file name to hyperlpr
from hyperlpr import pipline as pp
import cv2
# 自行修改文件名
image = cv2.imread("images/car5.jpg")
image, res = pp.SimpleRecognizePlate(image)
print('-------------')
print(res)
It's easy to implement. When running other codes on the Internet, it is really only necessary to take pictures once. It wastes a lot of time to make a summary here.