思路1:通过图像熵检测,“无内容”图像熵较小,可通过设置阈值检测“无内容”图像,计算图像熵可参考:https://www.cnblogs.com/niulang/p/12195152.html
思路2:检测图像中连通区域个数和面积
思路2代码:
import cv2
import numpy as np
import math
import time
import os
import shutil def get_cell_cnt(img_):
x, y = img_.shape[0:2]
img_ = cv2.resize(img_, (int(y/4), int(x/4)))
ret, binary = cv2.threshold(img_, 95, 255, cv2.THRESH_BINARY)
cv2.imwrite(path + 'abc.png', binary)
kernel_1 = np.ones((3,3),np.uint8) # 定义核
opening = cv2.morphologyEx(binary, cv2.MORPH_CLOSE, kernel_1)
cv2.imwrite(path + 'abc_1.png', opening)
opening_x = opening.shape[0]
opening_y = opening.shape[1]
opening[:,0] = 255
opening[:,opening_y-1] = 255
opening[0,:] = 255
opening[opening_x-1,:] = 255
# 检测图像连通区(输入为二值化图像)
image, contours, hierarchy = cv2.findContours(opening,1,2) # 检测连通区域
sign = 0
for n in range(len(contours)):
cnt = contours[n]
area = cv2.contourArea(cnt)
if area > 12:
sign = sign + 1
# cv2.imwrite(path + 'abc_open.png', opening)
# cv2.imwrite(path + 'abc_close_range.png', img_)
return sign
for path_ in open('list.txt'):
t1 = time.time()
path = path_[:-1]
image = cv2.imread(path,0)
t2 = time.time()
cell_cnt = get_cell_cnt(image)
t3 = time.time()
if cell_cnt < 2:
newpath = os.path.join('copy', path.split('/')[1])
shutil.copy(path, newpath)
print('time',t2-t1,t3-t1)