一、阈值处理理论
阈值处理指剔除图像内像素值高于一定值或低于一定值的像素点。如设阈值为127，然后：
- 将图像内所有像素值大于127的像素点的值设为255.
- 将图像内所有像素值小于127的像素点的值设为0.
二、普通阈值
def threshold_demo(image):
gray = cv.cvtColor(image, cv.COLOR_BGR2GRAY)
# 二值化阈值THRESH_BINARY：对于灰度值大于阈值thresh的像素点，将其灰度值设定为最大值；对于灰度值小于或等于阈值thresh的像素点，将其灰度值设定为0
# 反二值化阈值THRESH_BINARY_INV：对于灰度值大于阈值thresh的像素点，将其灰度值设定为0；对于灰度值小于或等于阈值thresh的像素点，将其灰度值设定为255
# 截断阈值化THRESH_TRUNC：对于像素值大于127的像素点，将其像素值设定为127；对于像素值小于或等于127的像素点，将其像素值将保持不变
# THRESH_OTSU方法：遍历当前图像所有可能阈值，从而找到最佳的阈值
"""
threshold(src, thresh, maxval, type, dst=None)：进行阈值化处理
src：要进行阈值分割的图像，可以是多通道的，8位或32位浮点型数值
thresh：设定的阈值
maxval：当type为THRESH_BINARY或 THRESH_BINARY_INV 类型时，需要设定的最大值
type：阈值要分割的类型
ret：返回的阈值
dst：阈值分割结果图像，与原始图像具有相同大小和类型
"""
ret, binary = cv.threshold(gray, 127, 255, cv.THRESH_BINARY|cv.THRESH_OTSU)
print('threshold value: %s' % ret)
cv.imshow('threshold_demo', binary)
三、自适应阈值
def local_demo(image):
gray = cv.cvtColor(image, cv.COLOR_BGR2GRAY)
"""
adaptiveThreshold(src, maxValue, adaptiveMethod, thresholdType, blockSize, C, dst=None)
src：处理的原始图像，必须是8位单通道的图像
maxValue：最大值
adaptiveMethod：自适应方法，逐个像素地计算自适应阈值。自适应阈值等于每个像素由参数blockSize所指定的加权平均值减去常量C
thresholdType：阈值处理方式，必须是THRESH_BINARY或 THRESH_BINARY_INV 中的一个
blockSize：块大小，表示一个像素在计算其阈值时所使用的领域尺寸，通常为3、5、7奇数等
C：常量
"""
# 自适应阈值ADAPTIVE_THRESH_MEAN_C：邻域所有像素点的权重值是一致的
# 高斯阈值ADAPTIVE_THRESH_GAUSSIAN_C：与邻域各个像素点到中心点的距离有关，通过高斯方程得到各个点的权重值。
binary = cv.adaptiveThreshold(gray, 255, cv.ADAPTIVE_THRESH_GAUSSIAN_C, cv.THRESH_BINARY, 25, 10)
cv.imshow('local_demo', binary)
四、自定义阈值
def custom_threshold(image):
gray = cv.cvtColor(image, cv.COLOR_BGR2GRAY)
h, w = gray.shape[:2]
# 变成一位数组
m = np.reshape(gray, [1, w*h])
print('一维数组m：', m)
mean = m.sum()/(w*h+6)
print('mean：', mean)
# 分割
ret, binary = cv.threshold(gray, mean, 255, cv.THRESH_BINARY)
cv.imshow('custom_threshold', binary)
五、超大图像普通阈值
def big_qimage_binary(image):
print(image.shape)
# 一、分块,指定宽高
cw = 256
ch = 256
h, w = image.shape[:2]
gray = cv.cvtColor(image, cv.COLOR_BGR2GRAY)
for row in range(0, h, ch):
for col in range(0, w, cw):
# 图像roi
roi = gray[row:row + ch, col:cw + col]
print(np.std(roi), np.mean(roi))
ret, dst = cv.threshold(roi, 0, 255, cv.THRESH_BINARY | cv.THRESH_OTSU)
# 空白图像过滤
gray[row:row + ch, col:cw + col] = dst
print(np.std(dst), np.mean(dst))
print()
cv.imwrite('D:/OpenCV/opencv-python/yang1.jpg', gray)
cv.imshow('big_qimage_binary', gray)
六、超大图像自适应阈值
def big_jimage_binary(image):
print(image.shape)
# 一、分块,指定宽高
cw = 256
ch = 256
h, w = image.shape[:2]
gray = cv.cvtColor(image, cv.COLOR_BGR2GRAY)
for row in range(0, h, ch):
for col in range(0, w, cw):
roi = gray[row:row + ch, col:cw + col]
dst = cv.adaptiveThreshold(roi, 255, cv.ADAPTIVE_THRESH_GAUSSIAN_C, cv.THRESH_BINARY, 127, 20)
gray[row:row + ch, col:cw + col] = dst
print(np.std(dst), np.mean(dst))
print()
cv.imwrite('D:/OpenCV/opencv-python/yang2.jpg', gray)
cv.imshow('big_jimage_binary', gray)
七、超大图像自定义形式
def big_zimage_binary(image):
print(image.shape)
cw = 256
ch = 256
h, w = image.shape[:2]
gray = cv.cvtColor(image, cv.COLOR_BGR2GRAY)
for row in range(0, h, ch):
for col in range(0, w, cw):
roi = gray[row:row + ch, col:cw + col]
# np.std(roi):计算沿指定轴的标准差。返回数组元素的标准差
dev = np.std(roi)
if dev < 15:
gray[row:row + ch, col:cw + col] = 255
else:
# dst = cv.adaptiveThreshold(roi, 255, cv.ADAPTIVE_THRESH_GAUSSIAN_C, cv.THRESH_BINARY, 127, 20)
ret, dst = cv.threshold(roi, 0, 255, cv.THRESH_BINARY | cv.THRESH_OTSU)
gray[row:row + ch, col:cw + col] = dst
print(np.std(dst), np.mean(dst))
print()
cv.imwrite('D:/OpenCV/opencv-python/yang4.jpg', gray)
cv.imshow('big_zimage_binary', gray)
