识别纸张边缘算法，运行clipper.py 可以达到以下的效果
把XXXXXX_clipper.py中的两个全局变量进行修改： 目标图片大小 TARGET_H = 672 TARGET_W = 1369
from skimage_clipper import main_img_preprocess
from skimage import io
origin_img = io.imread('your_img_path')
crop_img = main_img_preprocess(origin_img)
# you can save img by below code
# io.imsave('croped.jpg', crop_img)from opencv_clipper import main_img_preprocess
import cv2
origin_img = cv2.imread('your_img_path')
crop_img = main_img_preprocess(origin_img)
# you can save img by below code
# cv2.imwrite('croped.jpg', crop_img)同样的图片，opencv方式比skimage方式要快很多。 典型的（1400*700左右的图片）： opencv方式:100 ms skimage方式：1000 ms