基于深度学习车牌识别中遮挡车牌分类

最近实验室在做基于车牌识别系统的开发，发现检测后的车牌有一些仍然是污损车牌及没有车牌，为了使这部分车牌不进入识别系统，采用了cnn的分类系统对车牌进行划分，划分为遮挡车牌和非遮挡车牌，测试1000张车牌数据，分类准确率还不错，达到95%

框架：Tensorflow

方法：CNN卷积神经网络

数据集：2个数据集，如图，0文件夹储存完整清晰车牌数据，1文件夹存储污损车牌

测试集数据放置方式相同。

下面上训练代码

from skimage import io,transform
import glob
import os
import tensorflow as tf
import numpy as np
import time

#数据集地址
path='./car_train_data/'
#模型保存地址
model_path='./model/model.ckpt'
#基础学习率
LEARNING_RATE_BASE=0.001
#学习率的衰减率
LEARNING_RATE_DECAY=0.99
#滑动平均衰减率
MOVING_AVERAGE_DECAY=0.99
#将所有的图片resize成100*100
w=100
h=100
c=3
gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.7)

#读取图片
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):
        for im in glob.glob(folder+'/*.jpg'):
            print('reading the images:%s'%(im))
            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)
data,label=read_img(path)

#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, 2],
                                      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", [2], 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,True,regularizer)
batch_size=16
n_epoch=70
#(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')
#增加平均滑动和衰减学习率
# global_step=tf.Variable(0,trainable=False)
# variable_averages=tf.train.ExponentialMovingAverage(MOVING_AVERAGE_DECAY,global_step)
# variable_averages_op=variable_averages.apply(tf.trainable_variables())
# learning_rate=tf.train.exponential_decay(LEARNING_RATE_BASE,global_step,len(data)/batch_size,LEARNING_RATE_DECAY)

cross_entropy=tf.nn.sparse_softmax_cross_entropy_with_logits(logits=logits, labels=y_)
cross_entropy_mean=tf.reduce_mean(cross_entropy)
loss=cross_entropy_mean+tf.add_n(tf.get_collection('losses'))
# train_step=tf.train.GradientDescentOptimizer(learning_rate).minimize(loss,global_step=global_step)
train_op=tf.train.AdamOptimizer(learning_rate=0.001).minimize(loss)
# with tf.control_dependencies([train_step,variable_averages_op]):
#    train_op=tf.no_op(name='train')
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



saver=tf.train.Saver()
# sess=tf.Session()
sess = tf.Session(config=tf.ConfigProto(gpu_options=gpu_options))
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("step: %d" % (epoch + 1))
    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))
saver.save(sess,model_path)
sess.close()

训练过程中，加入正则化和dropout 方法进行训练，防止过拟合。数据集中，训练集和验证集比例为8：2

可以根据数据大小去更改batch_size，epoch值。

下面是测试代码

from skimage import io,transform
import tensorflow as tf
import numpy as np
import os
import types
import glob
from PIL import Image
import shutil
path='./car_test_data/'
gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.7)

car_dict = {0:'is7',1:'not7'}

w=100
h=100
c=3
imgs = []
images = []
labels = []
names = []
def read_one_image(path):

    cate = [path + x for x in os.listdir(path) if os.path.isdir(path + x)]

    for idx, folder in enumerate(cate):
        for im in glob.glob(folder + '/*.jpg'):
            data1 = io.imread(im)
            data1=transform.resize(data1, (w, h))
            images.append(im)
            names.append(im.split("/")[-1].split("\\")[-1])
            imgs.append(data1)
            labels.append(idx)
    return np.asarray(imgs, np.float32), np.asarray(labels, np.int32)
data,label=read_one_image(path)
#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]
with tf.Session(config=tf.ConfigProto(gpu_options=gpu_options)) as sess:

    saver = tf.train.import_meta_graph('./model/model.ckpt.meta')
    saver.restore(sess,tf.train.latest_checkpoint('./model/'))
    #设置batch_size
    batch_size=8
    graph = tf.get_default_graph()
    x = graph.get_tensor_by_name("x:0")
    y_ = graph.get_tensor_by_name("y_:0")
    feed_dict = {x:data,y_:label}
    logits = graph.get_tensor_by_name("logits_eval:0")
    n_batch = 0
    # for x_test, y_test in minibatches(data, label, batch_size, shuffle=True):
    #     print("The step ",n_batch+1,"to input images.")

        # n_batch=n_batch+1
    classification_result = sess.run(logits, feed_dict)
    correct_prediction = tf.equal(tf.cast(tf.argmax(logits, 1), tf.int32), y_)
    acc = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
    accuracy=sess.run(acc,feed_dict=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车的分类
    output = []
    output = tf.argmax(classification_result,1).eval()
    for i in range(len(output)):
        print("第",i+1,"张图车位数预测:"+car_dict[output[i]]+"实际: "+car_dict[labels[i]])
        # img=Image.open(images[i])
        print(names[i])
        if output[i]==0:
            shutil.copy(images[i],"./car_data/0/%s"%names[i])
           # img.save("./car_data/0/%s"%names[i])
        else:
            shutil.copy(images[i], "./car_data/1/%s" % names[i])
            # img.save("./car_data/1/%s"%names[i])
    print("accuracy=%g" % accuracy)

注意，训练代码和测试代码的数据都分为0和1，训练数据的0和1文件夹是放置项目里的car_train_data,测试数据的0和1文件夹是放置项目里的car_test_data里。

以上代码用的简单模型做的简单分类，效果还不错，如果还有更好的模型实现效果，也欢迎指出。