图像二值化、提取边缘

最新推荐文章于 2024-07-31 18:11:04 发布

fengjiexyb

最新推荐文章于 2024-07-31 18:11:04 发布

阅读量7.5k

点赞数

分类专栏： Python 图像识别

本文链接：https://blog.csdn.net/fengjiexyb/article/details/78031213

版权

这段代码用于LCD数字数据集的灰度和二值化处理。通过OpenCV库，应用高斯模糊、Otsu二值化等方法对图像进行预处理，同时展示了如何使用自适应阈值提取图像边缘。最后，将处理后的图像保存为JPEG格式。

摘要由CSDN通过智能技术生成

# -*- coding:utf-8 -*-
'''
@finish time:20170914
@author；fengjiexyb
this code is used for my lcd digit dataset gray and binarization
'''

import pickle
from svmutil import *
from PCV.tools import imtools
import numpy as np
import struct
import matplotlib.pyplot as plt
import random
import logging
import os
from PIL  import Image
import cv2
path='C:/Users/fengjiexyb/Desktop/ttt/'
files=os.listdir(path)
features=[]
label = []
for index in files:
    image=Image.open(path+index)
    imNew=np.array(image.resize((50,50)).convert('L'))
    thresh = 10
    maxValue = 255
    #binarization used ostu in three lower lines
    blur = cv2.GaussianBlur(imNew, (5, 5), 0)
    ret3, dst = cv2.threshold(blur, 0, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)
    Image.fromarray(dst).save('new' + index, 'jpeg')

    # 以下两行二值化的效果不如ostu
    # 具体参数参考ht