图片水印处理技术：opencv

公司项目新增需求：

爬虫抓取的封面图，需要做水印去除处理，因此需要封装一个服务，进行水印处理！想到了opencv库！

 

mac安装（失败）：

本人可能由于mac版本问题并未成功，会报如下异常

ImportError: dlopen(/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/cv2/cv2.cpython-37m-darwin.so, 2): Symbol not found: _clock_gettime
  Referenced from: /Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/cv2/.dylibs/libavutil.56.22.100.dylib (which was built for Mac OS X 10.12)
  Expected in: /usr/lib/libSystem.B.dylib
 in /Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/cv2/.dylibs/libavutil.56.22.100.dylib

linux服务器安装（成功）：

pip install opencv-python

如果出现问题：

yum install -y python-qt4

 

关于去水印代码：

参考：

https://www.cnblogs.com/dockers/p/9902202.html

https://blog.csdn.net/Maximun/article/details/85064274

https://www.jianshu.com/p/2da98431a779

https://blog.csdn.net/u011983997/article/details/80420339

图片：

def parse_image():
    path = "./demo1.jpg"  # 记得不要有中文路径
    img = cv2.imread(path)
    print(img)
    height, width = img.shape[0:2]
    # 开始操作
    thresh = cv2.inRange(img, np.array([0, 0, 0]), np.array([192, 192, 192]))
    scan = np.ones((3, 3), np.uint8)
    cor = cv2.dilate(thresh, scan, iterations=1)
    specular = cv2.inpaint(img, cor, 5, flags=cv2.INPAINT_TELEA)
    cv2.imwrite("./demo_result.png", specular)
    # 操作结束，下面开始是输出图片的代码
    # cv2.namedWindow("image", 0)
    # cv2.resizedWindow("image", int(width / 2), int(height / 2))
    # cv2.imshow("image", img)
    # cv2.namedWindow("modified", 0)
    # cv2.resizeWindow("modified", int(width / 2), int(height / 2))
    # cv2.imshow("modified", specular)
    # cv2.waitKey(0)
    # cv2.destroyAllWindows()
    pass