Python+OpenCV图像处理之超大图像二值化

（一）分块处理超大图像的二值化问题

python实现，包括全局阈值和局部阈值方法

import cv2
import numpy as np


__author__ = "boboa"


def big_image_demo(image):
    print(image.shape)
    # 分成小块，宽高
    cw = 256
    ch = 256
    h, w = image.shape[:2]
    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    for row in range(0, h, ch):
        for col in range(0, w, cw):
            roi = gray[row:row + ch, col:col + cw]  # 获取分块
            # ret, dst = cv2.threshold(roi,0,255,cv2.THRESH_BINARY|cv2.THRESH_OTSU)   #全局阈值
            dst = cv2.adaptiveThreshold(roi, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY, 127, 20) # 局部阈值
            gray[row:row + ch, col:col + cw] = dst  # 分块覆盖
            print(np.std(dst), np.mean(dst))
    cv2.imwrite("image/result_bigimage.jpg", gray)



if __name__ == "__main__":
    img = cv2.imread("image/big.jpg")
    cv2.namedWindow("input image", cv2.WINDOW_AUTOSIZE)
    cv2.imshow("input image", img)
    big_image_demo(img)
    cv2.waitKey(0)
    cv2.destroyAllWindows()

原图

全局阈值效果较差

 局部阈值效果较好

（二）图像ROI与空白区域过滤

python实现

import cv2
import numpy as np


__author__ = "boboa"


def big_image_demo(image):
    print(image.shape)
    # 分成小块，宽高
    cw = 256
    ch = 256
    h, w = image.shape[:2]
    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    for row in range(0, h, ch):
        for col in range(0, w, cw):
            roi = gray[row:row + ch, col:col + cw]  # 分块区域
            # 获取图像的标准差和平均值  
            print(np.std(roi), np.mean(roi))
            dev = np.std(roi)  # 计算像素的标准差 
            if dev < 15:
                # 如果像素的标准差小于某一阈值 我们就理解这个区域的图像变化不大,即黑白区间不明显，我们全部赋值为255
                gray[row:row + ch, col:col + cw] = 255
            else:
                ret, dst = cv2.threshold(roi, 0, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU)
                gray[row: row + ch, col: col + cw] = dst
    cv2.imwrite("image/result_bigimage.jpg", gray)

if __name__ == "__main__":
    img = cv2.imread("image/big.jpg")
    cv2.namedWindow("input image", cv2.WINDOW_AUTOSIZE)
    cv2.imshow("input image", img)
    big_image_demo(img)
    cv2.waitKey(0)
    cv2.destroyAllWindows()

运行结果，可以看出过滤之后，效果有所改善