【python】TensorFlow框架下CNN神经网络的花卉识别系统

提前说明一下，本文的CNN神经网络模型是参考网上诸多相关CNN图像分类大牛的博客修改的，在模型的基础上，用python的Flask框架搭载了一个web页面用来可视化展示。

第一步，爬取图片数据集

用python实现了一个非常简单的网络爬虫，对百度图片接口 http://image.baidu.com/search/acjson?tn=resultjson_com&ipn=rj&ct=201326592&is=&fp=result&queryWord=%E9%AB%98%E6%B8%85%E5%8A%A8%E6%BC%AB&cl=2&lm=-1&ie=utf-8&oe=utf-8&adpicid=&st=-1&z=&ic=0&word=%E4%BA%8C%E6%AC%A1%E5%85%83&s=&se=&tab=&width=&height=&face=0&istype=2&qc=&nc=1&fr=&pn=60&rn=30&gsm=1000000001e&1486375820481= 发送Http请求，返回Json串如下：

我们可以看到，data下的middleURL就是 我们想要的图片链接。于是，再向这个图片链接发请求，就可以获取到我们想要的图片了。代码如下：

# _*_ coding:utf-8 _*_
''''''
'''
    1.通过关键字进入图片界面
    
    2.加载图片
    queryWord:可爱图片
    word:可爱图片    
    pn:60
    gsm:3c
'''
import requests
import json
import time
import os

#要修改的参数列表
queryWord=input('请输入您要搜索的图片：')
pn=0
gsm=str(hex(pn))[-2:]
timestrp=int(time.time()*1000)
#num表示照片数量
num=1
#while实现类似翻页功能，遍历所有图片信息
while True:
    #请求的url
    url='https://image.baidu.com/search/acjson?' \
        'tn=resultjson_com&ipn=rj&ct=201326592&' \
        'is=&fp=result&queryWord={0}&cl=2&lm=-1&ie=utf-8&' \
        'oe=utf-8&adpicid=&st=-1&z=&ic=0&word={0}&s=&se=' \
        '&tab=&width=&height=&face=0&istype=2&qc=&nc=1&fr=&pn={1}&rn=30&gsm={2}&{3}='
    #伪装头部
    header={
        'User-Agent':'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/64.0.3282.119 Safari/537.36'
    }
    #解析为json()语句
    try:
        r_mus=requests.get(url.format(queryWord,pn,gsm,timestrp),headers=header).json()
    except BaseException as e:
        print("此处有错误%s"%e)
    print(r_mus)
    #遍历每一张图片信息
    for image in r_mus['data']:
        if image:
            #获取图片地址
            i_url=image['middleURL']
            #请求该地址
            r_img=requests.get(i_url,headers=header,stream=True).raw.read()
            print('正在读取第{}张图片'.format(num))
            num+=1
            time.sleep(0.7)
            #创建pictures目录
            if os.path.exists('data/other/'):
                pass
            else:
                os.mkdir('data/other/')
            #保存图片到文件夹pictures
            with open('data/other/'+str(int(time.time()))+'.jpg','wb')as files:
                files.write(r_img)
    listNum = r_mus['listNum']
    if listNum>pn:
        pn+=30
        gsm = str(hex(pn))[-2:]
        time.sleep(5)
    else:
        break

第二步，训练模型

模型借鉴的网上大佬 的博客模型。数据使用了一部分自己的数据集，一部分开源的花卉数据集。对模型进行训练，把训练好的模型放在model文件夹下
具体代码如下：

from skimage import io,transform
import glob
import os
import tensorflow as tf
import numpy as np
import time
import os
os.environ["CUDA_VISIBLE_DEVICES"] = "-1"

#读取花朵图片
def read_img(path):
    cate=[path+x for x in os.listdir(path) if os.path.isdir(path+x)]
    imgs=[]
    labels=[]
    for idx,folder in enumerate(cate):
        print('reading the dirs :%s' % (folder))
        for im in glob.glob(folder+'/*.jpg'):
            img=io.imread(im)
            img=transform.resize(img,(w,h))
            imgs.append(img)
            labels.append(idx)
    return np.asarray(imgs,np.float32),np.asarray(labels,np.int32)

def inference(input_tensor, train, regularizer):
    with tf.variable_scope('layer1-conv1'):
        conv1_weights = tf.get_variable("weight",[5,5,3,32],initializer=tf.truncated_normal_initializer(stddev=0.1))
        conv1_biases = tf.get_variable("bias", [32], initializer=tf.constant_initializer(0.0))
        conv1 = tf.nn.conv2d(input_tensor, conv1_weights, strides=[1, 1, 1, 1], padding='SAME')
        relu1 = tf.nn.relu(tf.nn.bias_add(conv1, conv1_biases))

    with tf.name_scope("layer2-pool1"):
        pool1 = tf.nn.max_pool(relu1, ksize = [1,2,2,1],strides=[1,2,2,1],padding="VALID")

    with tf.variable_scope("layer3-conv2"):
        conv2_weights = tf.get_variable("weight",[5,5,32,64],initializer=tf.truncated_normal_initializer(stddev=0.1))
        conv2_biases = tf.get_variable("bias", [64], initializer=tf.constant_initializer(0.0))
        conv2 = tf.nn.conv2d(pool1, conv2_weights, strides=[1, 1, 1, 1], padding='SAME')
        relu2 = tf.nn.relu(tf.nn.bias_add(conv2, conv2_biases))

    with tf.name_scope("layer4-pool2"):
        pool2 = tf.nn.max_pool(relu2, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID')

    with tf.variable_scope("layer5-conv3"):
        conv3_weights = tf.get_variable("weight",[3,3,64,128],initializer=tf.truncated_normal_initializer(stddev=0.1))
        conv3_biases = tf.get_variable("bias", [128], initializer=tf.constant_initializer(0.0))
        conv3 = tf.nn.conv2d(pool2, conv3_weights, strides=[1, 1, 1, 1], padding='SAME')
        relu3 = tf.nn.relu(tf.nn.bias_add(conv3, conv3_biases))

    with tf.name_scope("layer6-pool3"):
        pool3 = tf.nn.max_pool(relu3, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID')

    with tf.variable_scope("layer7-conv4"):
        conv4_weights = tf.get_variable("weight",[3,3,128,128],initializer=tf.truncated_normal_initializer(stddev=0.1))
        conv4_biases = tf.get_variable("bias", [128], initializer=tf.constant_initializer(0.0))
        conv4 = tf.nn.conv2d(pool3, conv4_weights, strides=[1, 1, 1, 1], padding='SAME')
        relu4 = tf.nn.relu(tf.nn.bias_add(conv4, conv4_biases))

    with tf.name_scope("layer8-pool4"):
        pool4 = tf.nn.max_pool(relu4, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID')
        nodes = 6*6*128
        reshaped = tf.reshape(pool4,[-1,nodes])

    with tf.variable_scope('layer9-fc1'):
        fc1_weights = tf.get_variable("weight", [nodes, 1024],
                                      initializer=tf.truncated_normal_initializer(stddev=0.1))
        if regularizer != None: tf.add_to_collection('losses', regularizer(fc1_weights))
        fc1_biases = tf.get_variable("bias", [1024], initializer=tf.constant_initializer(0.1))

        fc1 = tf.nn.relu(tf.matmul(reshaped, fc1_weights) + fc1_biases)
        if train: fc1 = tf.nn.dropout(fc1, 0.5)

    with tf.variable_scope('layer10-fc2'):
        fc2_weights = tf.get_variable("weight", [1024, 512],
                                      initializer=tf.truncated_normal_initializer(stddev=0.1))
        if regularizer != None: tf.add_to_collection('losses', regularizer(fc2_weights))
        fc2_biases = tf.get_variable("bias", [512], initializer=tf.constant_initializer(0.1))

        fc2 = tf.nn.relu(tf.matmul(fc1, fc2_weights) + fc2_biases)
        if train: fc2 = tf.nn.dropout(fc2, 0.5)

    with tf.variable_scope('layer11-fc3'):
        fc3_weights = tf.get_variable("weight", [512, 5],
                                      initializer=tf.truncated_normal_initializer(stddev=0.1))
        if regularizer != None: tf.add_to_collection('losses', regularizer(fc3_weights))
        fc3_biases = tf.get_variable("bias", [5], initializer=tf.constant_initializer(0.1))
        logit = tf.matmul(fc2, fc3_weights) + fc3_biases

    return logit

def minibatches(inputs=None, targets=None, batch_size=None, shuffle=False):
        assert len(inputs) == len(targets)
        if shuffle:
            indices = np.arange(len(inputs))
            np.random.shuffle(indices)
        for start_idx in range(0, len(inputs) - batch_size + 1, batch_size):
            if shuffle:
                excerpt = indices[start_idx:start_idx + batch_size]
            else:
                excerpt = slice(start_idx, start_idx + batch_size)
            yield inputs[excerpt], targets[excerpt]

def semantic_alignment(src_feature, tgt_feature, src_label, tgt_label, num_classes=2):
    '''
    input:
    src_feature: feature from source domain
    tgt_feature: feature from target somain
    src_label: source label(one-hot encoding)
    tgt_label: target label(one-hot encoding)
    num_classes : the number of class(e.g., 2)

    output:
    semantic_loss : the semantic loss between domains.
    '''

    source_result = tf.argmax(src_label, 1)  # source label
    target_result = tf.argmax(tgt_label, 1)  # target label
    ones = tf.ones_like(src_feature)#得到一个与源域数据格式一致的全1的张量
    print('ones',ones.shape)
    print('source_result', source_result.shape)
    print('target_result', target_result.shape)
    current_source_count = tf.unsorted_segment_sum(ones, source_result, num_classes)#计算出当前源域数据
    current_target_count = tf.unsorted_segment_sum(ones, target_result, num_classes)#计算出当前目标域数据
    current_positive_source_count = tf.maximum(current_source_count, tf.ones_like(current_source_count))#返回当前源域数据与之间的最大值
    current_positive_target_count = tf.maximum(current_target_count, tf.ones_like(current_target_count))#返回当前目标域数据与之间的最大值
    current_source_centroid = tf.divide(tf.unsorted_segment_sum(data=src_feature, segment_ids= \
    source_result, num_segments=num_classes), current_positive_source_count)
    current_target_centroid = tf.divide(tf.unsorted_segment_sum(data=tgt_feature, segment_ids= \
    target_result, num_segments=num_classes), current_positive_target_count)
    semantic_loss = tf.reduce_mean((tf.square(current_source_centroid - current_target_centroid)))
    return semantic_loss

if __name__ == '__main__':
    # 数据集地址
    path = 'D:/python/workspace/flower/data/flowers/'
    # 模型保存地址
    model_path = 'D:/python/workspace/flower/model/model.ckpt'
    #测试集地址
    path1= 'D:/python/workspace/flower/test1/'
    # 将所有的图片resize成100*100
    w = 100
    h = 100
    c = 3

    data, label = read_img(path)

    newlabel=[]
    # 打乱顺序
    num_example = data.shape[0]
    arr = np.arange(num_example)
    np.random.shuffle(arr)
    data = data[arr]
    label = label[arr]

    # 将所有数据分为训练集和验证集
    ratio = 0.8
    s = np.int(num_example * ratio)
    x_train = data[:s]
    y_train = label[:s]
    x_val = data[s:]
    y_val = label[s:]

    # -----------------构建网络----------------------
    # 占位符
    x = tf.placeholder(tf.float32, shape=[None, w, h, c], name='x')
    y_ = tf.placeholder(tf.int32, shape=[None, ], name='y_')


    def inference(input_tensor, train, regularizer):
        with tf.variable_scope('layer1-conv1'):
            conv1_weights = tf.get_variable("weight", [5, 5, 3, 32],
                                            initializer=tf.truncated_normal_initializer(stddev=0.1))
            conv1_biases = tf.get_variable("bias", [32], initializer=tf.constant_initializer(0.0))
            conv1 = tf.nn.conv2d(input_tensor, conv1_weights, strides=[1, 1, 1, 1], padding='SAME')
            relu1 = tf.nn.relu(tf.nn.bias_add(conv1, conv1_biases))

        with tf.name_scope("layer2-pool1"):
            pool1 = tf.nn.max_pool(relu1, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding="VALID")

        with tf.variable_scope("layer3-conv2"):
            conv2_weights = tf.get_variable("weight", [5, 5, 32, 64],
                                            initializer=tf.truncated_normal_initializer(stddev=0.1))
            conv2_biases = tf.get_variable("bias", [64], initializer=tf.constant_initializer(0.0))
            conv2 = tf.nn.conv2d(pool1, conv2_weights, strides=[1, 1, 1, 1], padding='SAME')
            relu2 = tf.nn.relu(tf.nn.bias_add(conv2, conv2_biases))

        with tf.name_scope("layer4-pool2"):
            pool2 = tf.nn.max_pool(relu2, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID')

        with tf.variable_scope("layer5-conv3"):
            conv3_weights = tf.get_variable("weight", [3, 3, 64, 128],
                                            initializer=tf.truncated_normal_initializer(stddev=0.1))
            conv3_biases = tf.get_variable("bias", [128], initializer=tf.constant_initializer(0.0))
            conv3 = tf.nn.conv2d(pool2, conv3_weights, strides=[1, 1, 1, 1], padding='SAME')
            relu3 = tf.nn.relu(tf.nn.bias_add(conv3, conv3_biases))

        with tf.name_scope("layer6-pool3"):
            pool3 = tf.nn.max_pool(relu3, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID')

        with tf.variable_scope("layer7-conv4"):
            conv4_weights = tf.get_variable("weight", [3, 3, 128, 128],
                                            initializer=tf.truncated_normal_initializer(stddev=0.1))
            conv4_biases = tf.get_variable("bias", [128], initializer=tf.constant_initializer(0.0))
            conv4 = tf.nn.conv2d(pool3, conv4_weights, strides=[1, 1, 1, 1], padding='SAME')
            relu4 = tf.nn.relu(tf.nn.bias_add(conv4, conv4_biases))

        with tf.name_scope("layer8-pool4"):
            pool4 = tf.nn.max_pool(relu4, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID')
            nodes = 6 * 6 * 128
            reshaped = tf.reshape(pool4, [-1, nodes])

        with tf.variable_scope('layer9-fc1'):
            fc1_weights = tf.get_variable("weight", [nodes, 1024],
                                          initializer=tf.truncated_normal_initializer(stddev=0.1))
            if regularizer != None: tf.add_to_collection('losses', regularizer(fc1_weights))
            fc1_biases = tf.get_variable("bias", [1024], initializer=tf.constant_initializer(0.1))

            fc1 = tf.nn.relu(tf.matmul(reshaped, fc1_weights) + fc1_biases)
            if train: fc1 = tf.nn.dropout(fc1, 0.5)

        with tf.variable_scope('layer10-fc2'):
            fc2_weights = tf.get_variable("weight", [1024, 512],
                                          initializer=tf.truncated_normal_initializer(stddev=0.1))
            if regularizer != None: tf.add_to_collection('losses', regularizer(fc2_weights))
            fc2_biases = tf.get_variable("bias", [512], initializer=tf.constant_initializer(0.1))

            fc2 = tf.nn.relu(tf.matmul(fc1, fc2_weights) + fc2_biases)
            if train: fc2 = tf.nn.dropout(fc2, 0.5)

        with tf.variable_scope('layer11-fc3'):
            fc3_weights = tf.get_variable("weight", [512, 5],
                                          initializer=tf.truncated_normal_initializer(stddev=0.1))
            if regularizer != None: tf.add_to_collection('losses', regularizer(fc3_weights))
            fc3_biases = tf.get_variable("bias", [5], initializer=tf.constant_initializer(0.1))
            logit = tf.matmul(fc2, fc3_weights) + fc3_biases

        return logit


    # ---------------------------网络结束---------------------------
    regularizer = tf.contrib.layers.l2_regularizer(0.0001)
    logits = inference(x, False, regularizer)

    # (小处理)将logits乘以1赋值给logits_eval，定义name，方便在后续调用模型时通过tensor名字调用输出tensor
    b = tf.constant(value=1, dtype=tf.float32)
    logits_eval = tf.multiply(logits, b, name='logits_eval')

    loss = tf.nn.sparse_softmax_cross_entropy_with_logits(logits=logits, labels=y_)
    train_op = tf.train.AdamOptimizer(learning_rate=0.001).minimize(loss)
    correct_prediction = tf.equal(tf.cast(tf.argmax(logits, 1), tf.int32), y_)
    acc = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))


    # 定义一个函数，按批次取数据
    def minibatches(inputs=None, targets=None, batch_size=None, shuffle=False):
        assert len(inputs) == len(targets)
        if shuffle:
            indices = np.arange(len(inputs))
            np.random.shuffle(indices)
        for start_idx in range(0, len(inputs) - batch_size + 1, batch_size):
            if shuffle:
                excerpt = indices[start_idx:start_idx + batch_size]
            else:
                excerpt = slice(start_idx, start_idx + batch_size)
            yield inputs[excerpt], targets[excerpt]


    # 训练和测试数据，可将n_epoch设置更大一些

    n_epoch = 10
    batch_size = 64
    saver = tf.train.Saver()
    sess = tf.Session()
    sess.run(tf.global_variables_initializer())
    for epoch in range(n_epoch):
        start_time = time.time()

        # training
        train_loss, train_acc, n_batch = 0, 0, 0
        for x_train_a, y_train_a in minibatches(x_train, y_train, batch_size, shuffle=True):
            _, err, ac = sess.run([train_op, loss, acc], feed_dict={x: x_train_a, y_: y_train_a})
            train_loss += err;
            train_acc += ac;
            n_batch += 1
        print("%d  epoch" % epoch)
        print("   train loss: %f" % (np.sum(train_loss) / n_batch))
        print("   train acc: %f" % (np.sum(train_acc) / n_batch))

        # validation
        val_loss, val_acc, n_batch = 0, 0, 0
        for x_val_a, y_val_a in minibatches(x_val, y_val, batch_size, shuffle=False):
            err, ac = sess.run([loss, acc], feed_dict={x: x_val_a, y_: y_val_a})
            val_loss += err;
            val_acc += ac;
            n_batch += 1
        print("   validation loss: %f" % (np.sum(val_loss) / n_batch))
        print("   validation acc: %f" % (np.sum(val_acc) / n_batch))
        print("============================================================ ")
    saver.save(sess, model_path)
    sess.close()

训练好后，我们可以看到model下已经有了具体的模型文件，代表训练成功。

训练成功后，我们用一个小程序来测试一下我们的模型

from skimage import io,transform
import tensorflow as tf
import numpy as np


path1 = "D:/python/workspace/flower/data/flowers/dandelion/8223968_6b51555d2f_n.jpg"
path2 = "D:/python/workspace/flower/data/other/1582514704.jpg"


flower_dict = {0:'flower',1:'other'}

w=100
h=100
c=3

def read_one_image(path):
    img = io.imread(path)
    img = transform.resize(img,(w,h))
    return np.asarray(img)

with tf.Session() as sess:
    data = []
    data1 = read_one_image(path1)
    data2 = read_one_image(path2)
    data.append(data1)
    data.append(data2)
    print(data1.shape)
    saver = tf.train.import_meta_graph('D:/python/workspace/flower/model1/model.ckpt.meta')
    saver.restore(sess,tf.train.latest_checkpoint('D:/python/workspace/flower/model1/'))

    graph = tf.get_default_graph()
    x = graph.get_tensor_by_name("x:0")
    feed_dict = {x:data}

    logits = graph.get_tensor_by_name("logits_eval:0")

    classification_result = sess.run(logits,feed_dict)

    #打印出预测矩阵
    print(classification_result)
    #打印出预测矩阵每一行最大值的索引
    print(tf.argmax(classification_result,1).eval())
    #根据索引通过字典对应花的分类
    output = []
    output = tf.argmax(classification_result,1).eval()
    for i in range(len(output)):
        print("第",i+1,"朵花预测:"+flower_dict[output[i]])

结果如下：

可以看出我们的模型是成功的，准确识别出了对应文件夹的图片数据。

第三步，准备Web界面

把准备好的web前端页面引入到templates文件夹中，然后使用flask 搭建web服务器。

然后写个接口，用来上传图片以及调用之前训练好的模型对花朵数据进行识别。

def getType(path):
    w = 100
    h = 100
    c = 3
    img = io.imread(path)
    data = []
    data.append(transform.resize(img,(w,h,3)))
    with tf.Session() as sess:

        saver = tf.train.import_meta_graph('D:/python/workspace/flower/model1/model.ckpt.meta')
        saver.restore(sess, tf.train.latest_checkpoint('D:/python/workspace/flower/model1/'))

        graph = tf.get_default_graph()
        x = graph.get_tensor_by_name("x:0")
        feed_dict = {x: data}

        logits = graph.get_tensor_by_name("logits_eval:0")

        classification_result = sess.run(logits, feed_dict)
        # 打印出预测矩阵每一行最大值的索引
        print(classification_result)

        output = tf.argmax(classification_result, 1).eval()
        if (output[0] == 1):
            return "不是花"
    tf.reset_default_graph()

    with tf.Session() as sess1:
        saver1 = tf.train.import_meta_graph('D:/python/workspace/flower/model/model.ckpt.meta')
        saver1.restore(sess1, tf.train.latest_checkpoint('D:/python/workspace/flower/model/'))
        graph = tf.get_default_graph()
        x = graph.get_tensor_by_name("x:0")
        feed_dict = {x: data}

        logits = graph.get_tensor_by_name("logits_eval:0")

        classification_result = sess1.run(logits, feed_dict)
        # 打印出预测矩阵每一行最大值的索引

        output = tf.argmax(classification_result, 1).eval()
    return flower_dict[output[0]]

@app.route('/upload',methods=['POST'])
def upload():
    file = request.files.get('file')

    type = getType(file)

    res = file.filename +    "，类型是：" + type
    return json.dumps(res, ensure_ascii=False)

上传一张图片测试一下，看反馈结果。

成功识别出了对应的图片信息，代表系统已经开发完成。

本系统以上传至本人Github ，如果可以帮助大家欢迎大家star，follow