Coffee_detection/Coffee_project_train.py

#coding=utf-8

import os
import numpy as np
import tensorflow as tf
import datetime
import cv2
import imutils
import time
import os.path

from xml.dom import minidom
from PIL import Image

# -------------------------------------------
extract_to = "C:\\Users\\User\\Desktop\\tfcoffebean\\allpic\\"
dataset_images = "C:\\Users\\User\\Desktop\\tfcoffebean\\allpic\\"
dataset_labels = "C:\\Users\\User\\Desktop\\tfcoffebean\\allpic\\"
folderCharacter = "/"  # \\ is for windows
xml_file = "C:\\Users\\User\\Desktop\\tfcoffebean\\xml_file.txt"
object_xml_file = "C:\\Users\\User\\Desktop\\tfcoffebean\\xml_object.txt"

def chkEnv():
    if not os.path.exists(extract_to):
        os.makedirs(extract_to)
        print("no {} folder, created.".format(extract_to))
    if (not os.path.exists(dataset_images)):
        print("There is no such folder {}".format(dataset_images))
        quit()
    if (not os.path.exists(dataset_labels)):
        print("There is no such folder {}".format(dataset_labels))
        quit()
    if (not os.path.exists(xml_file)):
        f = open(xml_file, 'w')
        f.close()
        print("There is no xml file in {} ,created.".format(xml_file))
    if (not os.path.exists(object_xml_file)):
        f = open(object_xml_file, 'w')
        f.close()
        print("There is no object xml file in {} ,created.".format(object_xml_file))

def getLabels(imgFile, xmlFile):
    labelXML = minidom.parse(xmlFile)
    labelName = []
    labelXmin = []
    labelYmin = []
    labelXmax = []
    labelYmax = []
    totalW = 0
    totalH = 0
    countLabels = 0

    tmpArrays = labelXML.getElementsByTagName("name")
    for elem in tmpArrays:
        labelName.append(str(elem.firstChild.data))
    tmpArrays = labelXML.getElementsByTagName("xmin")
    for elem in tmpArrays:
        labelXmin.append(int(elem.firstChild.data))
    tmpArrays = labelXML.getElementsByTagName("ymin")
    for elem in tmpArrays:
        labelYmin.append(int(elem.firstChild.data))
    tmpArrays = labelXML.getElementsByTagName("xmax")
    for elem in tmpArrays:
        labelXmax.append(int(elem.firstChild.data))
    tmpArrays = labelXML.getElementsByTagName("ymax")
    for elem in tmpArrays:
        labelYmax.append(int(elem.firstChild.data))

    return labelName, labelXmin, labelYmin, labelXmax, labelYmax

def write_lale_images(label, img, saveto, filename):
    writePath = extract_to + label
    print("WRITE:", writePath)
    if not os.path.exists(writePath):
        os.makedirs(writePath)

        cv2.imwrite(writePath + folderCharacter + filename, img)

def extract_lab_to_imgs():
    chkEnv()
    i = 0
    for file in os.listdir(dataset_images):
        filename, file_extension = os.path.splitext(file)
        file_extension = file_extension.lower()
        if (
                file_extension == ".jpg" or file_extension == ".jpeg" or file_extension == ".png" or file_extension == ".bmp"):
            print("Processing: ", dataset_images + file)
            if not os.path.exists(dataset_labels + filename + ".xml"):
                print("Cannot find the file {} for the image.".format(dataset_labels + filename + ".xml"))
            else:
                image_path = dataset_images + file
                xml_path = dataset_labels + filename + ".xml"
                labelName, labelXmin, labelYmin, labelXmax, labelYmax = getLabels(image_path, xml_path)
                orgImage = cv2.imread(image_path)
                image = orgImage.copy()
                for id, label in enumerate(labelName):
                    cv2.rectangle(image, (labelXmin[id], labelYmin[id]), (labelXmax[id], labelYmax[id]), (0, 255, 0), 2)
                    label_area = orgImage[labelYmin[id]:labelYmax[id], labelXmin[id]:labelXmax[id]]
                    label_img_filename = filename + "_" + str(id) + ".png"
                    write_lale_images(labelName[id], label_area, extract_to, label_img_filename)
                #cv2.imshow("Image", imutils.resize(image, width=700))
                #k = cv2.waitKey(1)

def turn_image():
    dir1 = 'C:\\Users\\User\\Desktop\\tfcoffebean\\allpic\\brokenbean\\'
    save_path1 = "C:\\Users\\User\\Desktop\\tfcoffebean\\brokenbean\\"

    dir2 = 'C:\\Users\\User\\Desktop\\tfcoffebean\\allpic\\peaberrybean\\'
    save_path2 = "C:\\Users\\User\\Desktop\\tfcoffebean\\peaberrybean\\"

    dir3 = 'C:\\Users\\User\\Desktop\\tfcoffebean\\allpic\\\shellbean\\'
    save_path3 = "C:\\Users\\User\\Desktop\\tfcoffebean\\\shellbean\\"

    dir4 = 'C:\\Users\\User\\Desktop\\tfcoffebean\\allpic\\\wormbean\\'
    save_path4 = "C:\\Users\\User\\Desktop\\tfcoffebean\\\wormbean\\"

    x = [dir1, dir2, dir3, dir4]
    y = [save_path1, save_path2, save_path3, save_path4]

    i = 0
    c = -1
    for dir in x:
        dirListing = os.listdir(dir)

        c = c + 1
        for filename in os.listdir(dir):
            img = Image.open(dir + "/" + filename)
            i = i + 1
            img_rotate0 = img.transpose(Image.FLIP_LEFT_RIGHT)
            img_rotate1 = img.rotate(90)
            img_rotate2 = img.rotate(180)
            img_rotate3 = img.rotate(270)

            temp = filename[0: filename.find('_')]
            temp2 = int(temp) + 1200000
            temp3 = str(temp2) + '_' + str(i) + filename[filename.find('_'):]
            img_rotate0.save(y[c] + temp3)
            temp3 = str(temp2) + '_' + str(i + len(dirListing)) + filename[filename.find('_'):]
            img_rotate1.save(y[c] + temp3)
            temp3 = str(temp2) + '_' + str(i + len(dirListing) + len(dirListing)) + filename[filename.find('_'):]
            img_rotate2.save(y[c] + temp3)
            temp3 = str(temp2) + '_' + str(i + len(dirListing) + len(dirListing) + len(dirListing)) + filename[
                                                                                                      filename.find(
                                                                                                          '_'):]
            img_rotate3.save(y[c] + temp3)
        print(y[c])
        print(len(dirListing))
        print('done')

def changename():
    start = datetime.datetime.now()
    path_name1 = 'C:\\Users\\User\\Desktop\\tfcoffebean\\brokenbean\\'
    path_name2 = 'C:\\Users\\User\\Desktop\\tfcoffebean\\peaberrybean\\'
    path_name3 = 'C:\\Users\\User\\Desktop\\tfcoffebean\\shellbean\\'
    path_name4 = 'C:\\Users\\User\\Desktop\\tfcoffebean\\wormbean\\'

    save_path = 'C:\\Users\\User\\Desktop\\tfcoffebean\\data\\'
    image_size = 150
    a = 1
    for item in os.listdir(path_name1):
        image_source = cv2.imread(path_name1 + item)
        image = cv2.resize(image_source, (image_size, image_size), 0, 0, cv2.INTER_LINEAR)
        cv2.imwrite(save_path + ('0' + '_' + str(a)) + ".png", image)
        # os.rename(os.path.join(path_name1,item),os.path.join(path_name1,('0'+'_'+str(a)+'.png')))#os.path.join(path_name,item)表示找到每个文件的绝对路径并进行拼接操作
        a += 1
    b = 1
    for item in os.listdir(path_name2):
        image_source = cv2.imread(path_name2 + item)
        image = cv2.resize(image_source, (image_size, image_size), 0, 0, cv2.INTER_LINEAR)
        cv2.imwrite(save_path + ('1' + '_' + str(b)) + ".png", image)
        # os.rename(os.path.join(path_name2,item),os.path.join(path_name2,('1'+'_'+str(b)+'.png')))#os.path.join(path_name,item)表示找到每个文件的绝对路径并进行拼接操作
        b += 1
    c = 1
    for item in os.listdir(path_name3):  #
        image_source = cv2.imread(path_name3 + item)
        image = cv2.resize(image_source, (image_size, image_size), 0, 0, cv2.INTER_LINEAR)
        cv2.imwrite(save_path + ('2' + '_' + str(c)) + ".png", image)
        # os.rename(os.path.join(path_name3,item),os.path.join(path_name3,('2'+'_'+str(c)+'.png')))#os.path.join(path_name,item)表示找到每个文件的绝对路径并进行拼接操作
        c += 1
    d = 1
    for item in os.listdir(path_name4):
        image_source = cv2.imread(path_name4 + item)
        image = cv2.resize(image_source, (image_size, image_size), 0, 0, cv2.INTER_LINEAR)
        cv2.imwrite(save_path + ('3' + '_' + str(d)) + ".png", image)
        # os.rename(os.path.join(path_name4,item),os.path.join(path_name4,('3'+'_'+str(d)+'.png')))#os.path.join(path_name,item)表示找到每个文件的绝对路径并进行拼接操作
        d += 1

    end = datetime.datetime.now()
    print("更名執行時間：", end - start)

def cnn():
    # data file
    data_dir = "data"

    # train or test
    train = True
    # model address
    model_path = "model/image_model"

    def read_data(data_dir):
        datas = []
        labels = []
        fpaths = []
        for fname in os.listdir(data_dir):
            fpath = os.path.join(data_dir, fname)
            fpaths.append(fpath)
            image = Image.open(fpath)
            data = np.array(image) / 255.0
            label = int(fname.split("_")[0])

            datas.append(data)
            labels.append(label)

        datas = np.array(datas)
        labels = np.array(labels)

        print("shape of datas: {}\tshape of labels: {}".format(datas.shape, labels.shape))
        return fpaths, datas, labels

    fpaths, datas, labels = read_data(data_dir)

    # num_classes = len(set(labels))
    num_classes = 4

    datas_placeholder = tf.placeholder(tf.float32, [None, 150, 150, 3])

    labels_placeholder = tf.placeholder(tf.int32, [None])

    dropout_placeholdr = tf.placeholder(tf.float32)

    conv0 = tf.layers.conv2d(datas_placeholder, 20, 5, activation=tf.nn.relu)
    pool0 = tf.layers.max_pooling2d(conv0, [2, 2], [2, 2])

    conv1 = tf.layers.conv2d(pool0, 40, 4, activation=tf.nn.relu)
    pool1 = tf.layers.max_pooling2d(conv1, [2, 2], [2, 2])

    flatten = tf.layers.flatten(pool1)

    fc = tf.layers.dense(flatten, 400, activation=tf.nn.relu)

    dropout_fc = tf.layers.dropout(fc, dropout_placeholdr)

    logits = tf.layers.dense(dropout_fc, num_classes)

    predicted_labels = tf.arg_max(logits, 1)

    losses = tf.nn.softmax_cross_entropy_with_logits(
        labels=tf.one_hot(labels_placeholder, num_classes),
        logits=logits
    )
    mean_loss = tf.reduce_mean(losses)

    optimizer = tf.train.AdamOptimizer(learning_rate=1e-2).minimize(losses)

    saver = tf.train.Saver()

    with tf.Session() as sess:

        if train:
            start = datetime.datetime.now()
            print("train mode")

            sess.run(tf.global_variables_initializer())

            train_feed_dict = {
                datas_placeholder: datas,
                labels_placeholder: labels,
                dropout_placeholdr: 0.25
            }
            for step in range(500):
                _, mean_loss_val = sess.run([optimizer, mean_loss], feed_dict=train_feed_dict)

                if step % 5 == 0:
                    print("step = {}\tmean loss = {}".format(step, mean_loss_val))
            saver.save(sess, model_path)
            print("train done，save model{}".format(model_path))
            end = datetime.datetime.now()
            print("訓練執行時間：", end - start)
        else:

            print("test mode")
            saver.restore(sess, model_path)
            print("{}reload model".format(model_path))

            label_name_dict = {
                0: "Brokenbeans",
                1: "Peaberry",
                2: "shellbean",
                3: "Worms"
            }
            test_feed_dict = {
                datas_placeholder: datas,
                labels_placeholder: labels,
                dropout_placeholdr: 0
            }
            predicted_labels_val = sess.run(predicted_labels, feed_dict=test_feed_dict)

            for fpath, real_label, predicted_label in zip(fpaths, labels, predicted_labels_val):
                # real_label_name = label_name_dict[real_label]
                predicted_label_name = label_name_dict[predicted_label]
                print("{}\t => {}".format(fpath, predicted_label_name))
            dirListing = os.listdir(data_dir)
            print(len(dirListing))


if __name__ == '__main__':
    start = datetime.datetime.now()
    extract_lab_to_imgs()
    end = datetime.datetime.now()
    print("xml轉換img執行時間：", end - start)
    start1 = datetime.datetime.now()
    turn_image()
    end1 = datetime.datetime.now()
    print("轉動圖片執行時間：", end1 - start1)
    changename()
    cnn()