sudo pip3 install --upgrade setuptoolssudo pip3 install scikit-image -i http://pypi.douban.com/simple/ --trusted-host pypi.douban.com//后面为镜像源加速
pip install --default-timeout=1000 opencv-python
//使用清华源安装
pip3 install opencv-python -i https://pypi.tuna.tsinghua.edu.cn/simple

# -*- coding: utf-8 -*-
from skimage import transform
import cv2
#aHash方法计算图片的指纹数据
img=cv2.imread('/Users/develop/imageanalysis/RPC框架.png')#读入图片
length = 8
height = 8
sum = 0.0
k = 0
grey = []
# 修改图像大小
dst=transform.resize(img, (length, height))
for i in range(length):
    for j in range(height):
        (b,g,r) = dst[i,j]
        # 计算灰度值
        grey.append(r*0.3+g*0.59+b*0.11)#浮点灰度值计算
        sum+=grey[k]
        k=k+1
average = sum/(length*height)#算出平均的灰度数据
fingerprint = []
for i in range(length*height):
    if grey[i] > average:
        fingerprint.append(1)#如果灰度数据大于平均值，填入元素1，否则填入元素0
    else:
        fingerprint.append(0)
print(fingerprint)

# -*- coding: utf-8 -*-
from skimage import transform,data
import matplotlib.pyplot as plt
from skimage import io
import cv2
img=cv2.imread('/Users/develop/imageanalysis/RPC框架.png')#读入图片
length = 9
height = 8
sum = 0.0
k = 0
grey = []
dst=transform.resize(img, (length, height))
#灰度图转化
for i in range(length):
    for j in range(height):
        (b,g,r) = dst[i,j]
        grey.append(r*0.3+g*0.59+b*0.11)#浮点灰度值计算
        sum+=grey[k]
        k=k+1
# 获取每行的差值，为64位
differ = []
for i in range(length-1):#i 8
    for j in range(height):#j 8
        differ.append(grey[j*8+i+1]-grey[j*8+i])
#获取01值
fingerprint = []
for i in range((length-1)*height):#8*8
    if differ[i] > 0:
        fingerprint.append(1)
    else:
        fingerprint.append(0)
print(fingerprint)

# -*- coding: utf-8 -*-
import cv2
from skimage import transform
import numpy as np
#利用phash计算图像指纹
img=cv2.imread('/Users/develop/imageanalysis/RPC框架.png')#读入图片
length = 32
s_length = 8
dst=transform.resize(img, (length, length))
#print(type(dst))
#all 代表全图，建议32*32
matrix_all_grey = [[0 for i in range(length)] for i in range(length)]
#small 代表最低频率，8*8大小
matrix_small_grey = [[0 for i in range(s_length)] for i in range(s_length)]
# 灰度图转化
k = 0
for i in range(length):
    for j in range(length):
        (b,g,r) = dst[i,j]
        matrix_all_grey[i][j] = ( b + g + r )/3
# 缩小
fft_matrix_grey = np.fft.fft(matrix_all_grey)
for i in range(s_length):
    for j in range(s_length):
        matrix_small_grey[i][j] = fft_matrix_grey[i][j]
# 计算平均
sum_pixel = 0
for i in range(s_length):
    for j in range(s_length):
        sum_pixel += matrix_small_grey[i][j]
average = sum_pixel / (s_length*s_length)
bite = []
#赋予01值
for i in range(s_length):
    for j in range(s_length):
        if matrix_small_grey[i][j] > average :
            bite.append(1)
        else:
            bite.append(0)
print(bite)

split_count = 8  # 每个64位的phash值分为八段，每段8位
def split(key, split_count):
    pre_length = 64 / split_count
    return [key[i * pre_length: (i + 1) * pre_length] for i in range(split_count)]
class ImageManager(object):
    def __init__(self):
        self.phash = pHash() # pHash类
        self.phash_cache = [defaultdict(list) for i in range(split_count)] #
        self.init_phash_map()
    def init_phash_map(self):
        #把所有的phash存在sqlite里边，这边取出所有的Image
        for image in Image.select():
            self.add_to_image_cache(image)
    def add_to_image_cache(self, image):
        # 将hash值分割为8段
        key_split = split(bin(int(image.phash))[2:].rjust(64, '0'), split_count)
        for index, k in enumerate(key_split):
            self.phash_cache[index][k].append(image)
    def has_same(self, ori_image):
        phash = ori_image.phash
        key_split = split(bin(int(phash))[2:].rjust(64, '0'), split_count)
        result = set()
        for index, k in enumerate(key_split):
            if k in self.phash_cache[index]:
                for image in self.phash_cache[index][k]:
                    distance = self.distance(int(phash), int(image.phash))
                    if distance < 5 and ori_image.key != image.key:
                        result.add(image)
        if result:
            return True,list(result)
        return False,[]