本文實(shí)例為大家分享了python多進(jìn)程讀圖提取特征存npy的具體代碼,供大家參考,具體內(nèi)容如下
import multiprocessing
import os, time, random
import numpy as np
import cv2
import os
import sys
from time import ctime
import tensorflow as tf
image_dir = r"D:/sxl/處理圖片/漢字分類/train10/" #圖像文件夾路徑
data_type = 'test'
save_path = r'E:/sxl_Programs/Python/CNN/npy/' #存儲(chǔ)路徑
data_name = 'Img10' #npy文件名
char_set = np.array(os.listdir(image_dir)) #文件夾名稱列表
np.save(save_path+'ImgShuZi10.npy',char_set) #文件夾名稱列表
char_set_n = len(char_set) #文件夾列表長(zhǎng)度
read_process_n = 1 #進(jìn)程數(shù)
repate_n = 4 #隨機(jī)移動(dòng)次數(shù)
data_size = 1000000 #1個(gè)npy大小
shuffled = True #是否打亂
#可以讀取帶中文路徑的圖
def cv_imread(file_path,type=0):
cv_img=cv2.imdecode(np.fromfile(file_path,dtype=np.uint8),-1)
# print(file_path)
# print(cv_img.shape)
# print(len(cv_img.shape))
if(type==0):
if(len(cv_img.shape)==3):
cv_img = cv2.cvtColor(cv_img, cv2.COLOR_BGR2GRAY)
return cv_img
#多個(gè)數(shù)組按同一規(guī)則打亂數(shù)據(jù)
def ShuffledData(features,labels):
'''
@description:隨機(jī)打亂數(shù)據(jù)與標(biāo)簽,但保持?jǐn)?shù)據(jù)與標(biāo)簽一一對(duì)應(yīng)
'''
permutation = np.random.permutation(features.shape[0])
shuffled_features = features[permutation,:] #多維
shuffled_labels = labels[permutation] #1維
return shuffled_features,shuffled_labels
#函數(shù)功能:簡(jiǎn)單網(wǎng)格
#函數(shù)要求:1.無(wú)關(guān)圖像大小;2.輸入圖像默認(rèn)為灰度圖;3.參數(shù)只有輸入圖像
#返回?cái)?shù)據(jù):1x64*64維特征
def GetFeature(image):
#圖像大小歸一化
image = cv2.resize(image,(64,64))
img_h = image.shape[0]
img_w = image.shape[1]
#定義特征向量
feature = np.zeros(img_h*img_w,dtype=np.int16)
for h in range(img_h):
for w in range(img_w):
feature[h*img_h+w] = image[h,w]
return feature
# 寫數(shù)據(jù)進(jìn)程執(zhí)行的代碼:
def read_image_to_queue(queue):
print('Process to write: %s' % os.getpid())
for j,dirname in enumerate(char_set): # dirname 是文件夾名稱
label = np.where(char_set==dirname)[0][0] #文件夾名稱對(duì)應(yīng)的下標(biāo)序號(hào)
print('序號(hào):'+str(j),'讀 '+dirname+' 文件夾...時(shí)間:',ctime() )
for parent,_,filenames in os.walk(os.path.join(image_dir,dirname)):
for filename in filenames:
if(filename[-4:]!='.jpg'):
continue
image = cv_imread(os.path.join(parent,filename),0)
# cv2.imshow(dirname,image)
# cv2.waitKey(0)
queue.put((image,label))
for i in range(read_process_n):
queue.put((None,-1))
print('讀圖結(jié)束!')
return True
# 讀數(shù)據(jù)進(jìn)程執(zhí)行的代碼:
def extract_feature(queue,lock,count):
'''
@description:從隊(duì)列中取出圖片進(jìn)行特征提取
@queue:先進(jìn)先出隊(duì)列
lock:鎖,在計(jì)數(shù)時(shí)上鎖,防止沖突
count:計(jì)數(shù)
'''
print('Process %s start reading...' % os.getpid())
global data_n
features = [] #存放提取到的特征
labels = [] #存放標(biāo)簽
flag = True #標(biāo)志著進(jìn)程是否結(jié)束
while flag:
image,label = queue.get() #從隊(duì)列中獲取圖像和標(biāo)簽
if len(features) >= data_size or label == -1: #特征數(shù)組的長(zhǎng)度大于指定長(zhǎng)度,則開始存儲(chǔ)
array_features = np.array(features) #轉(zhuǎn)換成數(shù)組
array_labels = np.array(labels)
array_features,array_labels = ShuffledData(array_features,array_labels) #打亂數(shù)據(jù)
lock.acquire() # 鎖開始
# 拆分?jǐn)?shù)據(jù)為訓(xùn)練集,測(cè)試集
split_x = int(array_features.shape[0] * 0.8)
train_data, test_data = np.split(array_features, [split_x], axis=0) # 拆分特征數(shù)據(jù)集
train_labels, test_labels = np.split(array_labels, [split_x], axis=0) # 拆分標(biāo)簽數(shù)據(jù)集
count.value += 1 #下標(biāo)計(jì)數(shù)加1
str_features_name_train = data_name+'_features_train_'+str(count.value)+'.npy'
str_labels_name_train = data_name+'_labels_train_'+str(count.value)+'.npy'
str_features_name_test = data_name+'_features_test_'+str(count.value)+'.npy'
str_labels_name_test = data_name+'_labels_test_'+str(count.value)+'.npy'
lock.release() # 鎖釋放
np.save(save_path+str_features_name_train,train_data)
np.save(save_path+str_labels_name_train,train_labels)
np.save(save_path+str_features_name_test,test_data)
np.save(save_path+str_labels_name_test,test_labels)
print(os.getpid(),'save:',str_features_name_train)
print(os.getpid(),'save:',str_labels_name_train)
print(os.getpid(),'save:',str_features_name_test)
print(os.getpid(),'save:',str_labels_name_test)
features.clear()
labels.clear()
if label == -1:
break
# 獲取特征向量,傳入灰度圖
feature = GetFeature(image)
features.append(feature)
labels.append(label)
# # 隨機(jī)移動(dòng)4次
# for itime in range(repate_n):
# rMovedImage = randomMoveImage(image)
# feature = SimpleGridFeature(rMovedImage) # 簡(jiǎn)單網(wǎng)格
# features.append(feature)
# labels.append(label)
print('Process %s is done!' % os.getpid())
if __name__=='__main__':
time_start = time.time() # 開始計(jì)時(shí)
# 父進(jìn)程創(chuàng)建Queue,并傳給各個(gè)子進(jìn)程:
image_queue = multiprocessing.Queue(maxsize=1000) #隊(duì)列
lock = multiprocessing.Lock() #鎖
count = multiprocessing.Value('i',0) #計(jì)數(shù)
#將圖寫入隊(duì)列進(jìn)程
write_sub_process = multiprocessing.Process(target=read_image_to_queue, args=(image_queue,))
read_sub_processes = [] #讀圖子線程
for i in range(read_process_n):
read_sub_processes.append(
multiprocessing.Process(target=extract_feature, args=(image_queue,lock,count))
)
# 啟動(dòng)子進(jìn)程pw,寫入:
write_sub_process.start()
# 啟動(dòng)子進(jìn)程pr,讀取:
for p in read_sub_processes:
p.start()
# 等待進(jìn)程結(jié)束:
write_sub_process.join()
for p in read_sub_processes:
p.join()
time_end=time.time()
time_h=(time_end-time_start)/3600
print('用時(shí):%.6f 小時(shí)'% time_h)
print ("讀圖提取特征存npy,運(yùn)行結(jié)束!")
以上就是本文的全部?jī)?nèi)容,希望對(duì)大家的學(xué)習(xí)有所幫助,也希望大家多多支持腳本之家。
更多文章、技術(shù)交流、商務(wù)合作、聯(lián)系博主
微信掃碼或搜索:z360901061
微信掃一掃加我為好友
QQ號(hào)聯(lián)系: 360901061
您的支持是博主寫作最大的動(dòng)力,如果您喜歡我的文章,感覺我的文章對(duì)您有幫助,請(qǐng)用微信掃描下面二維碼支持博主2元、5元、10元、20元等您想捐的金額吧,狠狠點(diǎn)擊下面給點(diǎn)支持吧,站長(zhǎng)非常感激您!手機(jī)微信長(zhǎng)按不能支付解決辦法:請(qǐng)將微信支付二維碼保存到相冊(cè),切換到微信,然后點(diǎn)擊微信右上角掃一掃功能,選擇支付二維碼完成支付。
【本文對(duì)您有幫助就好】元

