求能用python实现基于概率分布的超声图像与自然图像性质差异分析这一课题的完整详细代码，其中超声图像所在的文件夹路径是'D:/zzz/us5'，自然图像所在的文件夹路径是'D:/zzz/na4'，两种图像的类型都是jpg,两种图像都有84张，两种图像的形状都相同，对比的特征需要包含颜色，形态和纹理，最后利用K-S检验来比较两种图像的概率密度函数

首先，需要安装必要的库，包括numpy、matplotlib、cv2和scipy:

import numpy as np 
import matplotlib.pyplot as plt 
import cv2 
from scipy.stats import ks_2samp

接下来，读入超声图像和自然图像，并将它们转化为灰度图像:

us_path = 'D:/zzz/us5/'
na_path = 'D:/zzz/na4/'

us_images = []
na_images = []

for i in range(1, 85):
    us_image = cv2.imread(us_path + str(i) + '.jpg', cv2.IMREAD_GRAYSCALE)
    na_image = cv2.imread(na_path + str(i) + '.jpg', cv2.IMREAD_GRAYSCALE)
    
    us_images.append(us_image)
    na_images.append(na_image)

然后，使用opencv中的函数计算图像的颜色、形态和纹理特征:

# 灰度直方图
def gray_hist(image):
    hist = cv2.calcHist([image],[0],None,[256],[0,256])
    return hist.flatten()

# 形态学特征
def morphology(image):
    kernel = cv2.getStructuringElement(cv2.MORPH_RECT,(3,3))
    opened = cv2.morphologyEx(image, cv2.MORPH_OPEN, kernel)
    closed = cv2.morphologyEx(opened, cv2.MORPH_CLOSE, kernel)
    return closed

# 纹理特征
def texture(image):
    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    height, width = gray.shape
    dst = np.zeros((height, width), np.uint8)
    for i in range(0, height-2):
        for j in range(0, width-2):
            x1 = gray[i, j]
            x2 = gray[i, j+1]
            x3 = gray[i, j+2]
            x4 = gray[i+1, j]
            x5 = gray[i+1, j+1]
            x6 = gray[i+1, j+2]
            x7 = gray[i+2, j]
            x8 = gray[i+2, j+1]
            x9 = gray[i+2, j+2]
            T = (int(x1)+int(x2)+int(x3)+int(x4)+int(x5)+int(x6)+int(x7)+int(x8)+int(x9))/9
            if x1 > T:
                a1 = 1
            else:
                a1 = 0
            if x2 > T:
                a2 = 2
            else:
                a2 = 0
            if x3 > T:
                a3 = 4
            else:
                a3 = 0
            if x4 > T:
                a4 = 8
            else:
                a4 = 0
            if x5 > T:
                a5 = 16
            else:
                a5 = 0
            if x6 > T:
                a6 = 32
            else:
                a6 = 0
            if x7 > T:
                a7 = 64
            else:
                a7 = 0
            if x8 > T:
                a8 = 128
            else:
                a8 = 0
            if x9 > T:
                a9 = 256
            else:
                a9 = 0
            sum = a1 + a2 + a3 + a4 + a5 + a6 + a7 + a8 + a9
            dst[i, j] = sum
    hist = cv2.calcHist([dst],[0],None,[256],[0,256])
    return hist.flatten()
    
us_color, us_shape, us_texture = [], [], []
na_color, na_shape, na_texture = [], [], []

for i in range(84):
    us_color.append(gray_hist(cv2.imread(us_path + str(i+1) + '.jpg')))
    na_color.append(gray_hist(cv2.imread(na_path + str(i+1) + '.jpg')))
    
    us_shape.append(gray_hist(morphology(cv2.imread(us_path + str(i+1) + '.jpg'))))
    na_shape.append(gray_hist(morphology(cv2.imread(na_path + str(i+1) + '.jpg'))))
    
    us_texture.append(texture(cv2.imread(us_path + str(i+1) + '.jpg')))
    na_texture.append(texture(cv2.imread(na_path + str(i+1) + '.jpg')))

最后，使用K-S检验来比较超声图像和自然图像的概率密度函数:

color_pvalue = ks_2samp(us_color, na_color)[1]
shape_pvalue = ks_2samp(us_shape, na_shape)[1]
texture_pvalue = ks_2samp(us_texture, na_texture)[1]

print('颜色差异检验p值:', color_pvalue)
print('形态差异检验p值:', shape_pvalue)
print('纹理差异检验p值:', texture_pvalue)

完整代码如下：

import numpy as np 
import matplotlib.pyplot as plt 
import cv2 
from scipy.stats import ks_2samp

us_path = 'D:/zzz/us5/'
na_path = 'D:/zzz/na4/'

us_images = []
na_images = []

for i in range(1, 85):
    us_image = cv2.imread(us_path + str(i) + '.jpg', cv2.IMREAD_GRAYSCALE)
    na_image = cv2.imread(na_path + str(i) + '.jpg', cv2.IMREAD_GRAYSCALE)
    
    us_images.append(us_image)
    na_images.append(na_image)
    
# 灰度直方图
def gray_hist(image):
    hist = cv2.calcHist([image],[0],None,[256],[0,256])
    return hist.flatten()

# 形态学特征
def morphology(image):
    kernel = cv2.getStructuringElement(cv2.MORPH_RECT,(3,3))
    opened = cv2.morphologyEx(image, cv2.MORPH_OPEN, kernel)
    closed = cv2.morphologyEx(opened, cv2.MORPH_CLOSE, kernel)
    return closed

# 纹理特征
def texture(image):
    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    height, width = gray.shape
    dst = np.zeros((height, width), np.uint8)
    for i in range(0, height-2):
        for j in range(0, width-2):
            x1 = gray[i, j]
            x2 = gray[i, j+1]
            x3 = gray[i, j+2]
            x4 = gray[i+1, j]
            x5 = gray[i+1, j+1]
            x6 = gray[i+1, j+2]
            x7 = gray[i+2, j]
            x8 = gray[i+2, j+1]
            x9 = gray[i+2, j+2]
            T = (int(x1)+int(x2)+int(x3)+int(x4)+int(x5)+int(x6)+int(x7)+int(x8)+int(x9))/9
            if x1 > T:
                a1 = 1
            else:
                a1 = 0
            if x2 > T:
                a2 = 2
            else:
                a2 = 0
            if x3 > T:
                a3 = 4
            else:
                a3 = 0
            if x4 > T:
                a4 = 8
            else:
                a4 = 0
            if x5 > T:
                a5 = 16
            else:
                a5 = 0
            if x6 > T:
                a6 = 32
            else:
                a6 = 0
            if x7 > T:
                a7 = 64
            else:
                a7 = 0
            if x8 > T:
                a8 = 128
            else:
                a8 = 0
            if x9 > T:
                a9 = 256
            else:
                a9 = 0
            sum = a1 + a2 + a3 + a4 + a5 + a6 + a7 + a8 + a9
            dst[i, j] = sum
    hist = cv2.calcHist([dst],[0],None,[256],[0,256])
    return hist.flatten()
    
us_color, us_shape, us_texture = [], [], []
na_color, na_shape, na_texture = [], [], []

for i in range(84):
    us_color.append(gray_hist(cv2.imread(us_path + str(i+1) + '.jpg')))
    na_color.append(gray_hist(cv2.imread(na_path + str(i+1) + '.jpg')))
    
    us_shape.append(gray_hist(morphology(cv2.imread(us_path + str(i+1) + '.jpg'))))
    na_shape.append(gray_hist(morphology(cv2.imread(na_path + str(i+1) + '.jpg'))))
    
    us_texture.append(texture(cv2.imread(us_path + str(i+1) + '.jpg')))
    na_texture.append(texture(cv2.imread(na_path + str(i+1) + '.jpg')))
    
color_pvalue = ks_2samp(us_color, na_color)[1]
shape_pvalue = ks_2samp(us_shape, na_shape)[1]
texture_pvalue = ks_2samp(us_texture, na_texture)[1]

print('颜色差异检验p值:', color_pvalue)
print('形态差异检验p值:', shape_pvalue)
print('纹理差异检验p值:', texture_pvalue)