前言

上一篇文章 神奇的目標檢測 已經介紹了目標檢測的基礎啦，目標檢測呢，就是在圖片中定位出目標的位置，把它“框”出來就好了，本篇文章使用Yolov3 和Yolov3-tiny，以訓練VOC2007和口罩檢測為例，教大家如何快速的搭建自己的目標檢測平臺，下面是資源鏈接：

一、Yolov3 和 Yolov3-tiny

2018 年，推出了Yolov3，相比于Yolov2 最主要的改進有，一就是加深了網路，使用53層Darknet，二就是采用了FPN（空間金字塔結構），三是使用了focal loss，對于tiny版本來說，只使用了簡單的44層卷積用作普通的特征提取，每個網格點使用兩個anchor boxes（Yolov3三個），所以tiny版本檢測速度是很快的哦~，但檢測效果就…，

1.網路結構

網路結構中包含了很多基礎塊，我們先實作這些基本的塊，然后像搭積木一樣將這些塊給組裝起來，每個塊的用途我已經寫在代碼注釋里了，

 #定義的卷積設定初始化方法和卷積步長和填充方式
@wraps(Conv2D)
def DarknetConv2D(*args, **kwargs):
    """Wrapper to set Darknet parameters for Convolution2D."""
    #定義卷積塊
    darknet_conv_kwargs = {'kernel_regularizer': l2(5e-4)}
    darknet_conv_kwargs['padding'] = 'valid' if kwargs.get('strides')==(2,2) else 'same'
    darknet_conv_kwargs.update(kwargs)
    return Conv2D(*args, **darknet_conv_kwargs)

def DarknetConv2D_BN_Leaky(*args, **kwargs):
    #定義的卷積塊包含了BN Leaky 激活函式
    """Darknet Convolution2D followed by BatchNormalization and LeakyReLU."""
    no_bias_kwargs = {'use_bias': False}
    no_bias_kwargs.update(kwargs)
    return compose(
        DarknetConv2D(*args, **no_bias_kwargs),
        BatchNormalization(),
        LeakyReLU(alpha=0.1))

def resblock_body(x, num_filters, num_blocks):
    '''A series of resblocks starting with a downsampling Convolution2D'''
    #定義 yolo 主干使用的殘差快
    # Darknet uses left and top padding instead of 'same' mode
    x = ZeroPadding2D(((1,0),(1,0)))(x)
    x = DarknetConv2D_BN_Leaky(num_filters, (3,3), strides=(2,2))(x)
    for i in range(num_blocks):
        y = compose(
                DarknetConv2D_BN_Leaky(num_filters//2, (1,1)),
                DarknetConv2D_BN_Leaky(num_filters, (3,3)))(x)
        x = Add()([x,y])
    return x

def darknet_body(x):
    '''Darknent body having 52 Convolution2D layers'''
    #darknet 53
    #卷積核大小3x3 32 個卷積核
    x = DarknetConv2D_BN_Leaky(32, (3,3))(x)
    x = resblock_body(x, 64, 1)
    x = resblock_body(x, 128, 2)
    x = resblock_body(x, 256, 8)
    x = resblock_body(x, 512, 8)
    x = resblock_body(x, 1024, 4)
    return x

def make_last_layers(x, num_filters, out_filters):
    '''6 Conv2D_BN_Leaky layers followed by a Conv2D_linear layer'''
    # 這里是輸入yolo，制造最后一層的代碼 也就是yolo head
    x = compose(
            DarknetConv2D_BN_Leaky(num_filters, (1,1)),
            DarknetConv2D_BN_Leaky(num_filters*2, (3,3)),
            DarknetConv2D_BN_Leaky(num_filters, (1,1)),
            DarknetConv2D_BN_Leaky(num_filters*2, (3,3)),
            DarknetConv2D_BN_Leaky(num_filters, (1,1)))(x)
    y = compose(
            DarknetConv2D_BN_Leaky(num_filters*2, (3,3)),
            DarknetConv2D(out_filters, (1,1)))(x)
    return x, y

yolov3-tiny

tiny 版本網路結構比較簡單，我們先來看一個圖：

網路中就是普通的卷積核和池化，且網路很淺，網路的計算程序如箭頭所示，是不是網路很簡單呀~~~~~，
我們接下來看代碼如何實作，

def tiny_yolo_body(inputs, num_anchors, num_classes):
	#-------------------------------------------------------------------
	# inputs 輸入向量  num_anchors anchor boxes的數量 num_classes 類別數
	#------------------------------------------------------------------
    '''Create Tiny YOLO_v3 model CNN body in keras.'''
    x1 = compose(
            DarknetConv2D_BN_Leaky(16, (3,3)),
            MaxPooling2D(pool_size=(2,2), strides=(2,2), padding='same'),
            DarknetConv2D_BN_Leaky(32, (3,3)),
            MaxPooling2D(pool_size=(2,2), strides=(2,2), padding='same'),
            DarknetConv2D_BN_Leaky(64, (3,3)),
            MaxPooling2D(pool_size=(2,2), strides=(2,2), padding='same'),
            DarknetConv2D_BN_Leaky(128, (3,3)),
            MaxPooling2D(pool_size=(2,2), strides=(2,2), padding='same'),
            DarknetConv2D_BN_Leaky(256, (3,3)))(inputs)
    x2 = compose(
            MaxPooling2D(pool_size=(2,2), strides=(2,2), padding='same'),
            DarknetConv2D_BN_Leaky(512, (3,3)),
            MaxPooling2D(pool_size=(2,2), strides=(1,1), padding='same'),
            DarknetConv2D_BN_Leaky(1024, (3,3)),
            DarknetConv2D_BN_Leaky(256, (1,1)))(x1)
    y1 = compose(
            DarknetConv2D_BN_Leaky(512, (3,3)),
            DarknetConv2D(num_anchors*(num_classes+5), (1,1)))(x2)

    x2 = compose(
            DarknetConv2D_BN_Leaky(128, (1,1)),
            UpSampling2D(2))(x2)
    y2 = compose(
            Concatenate(),
            DarknetConv2D_BN_Leaky(256, (3,3)),
            DarknetConv2D(num_anchors*(num_classes+5), (1,1)))([x2,x1])

    return Model(inputs, [y1,y2])

DarknetConv2D DarknetConv2D_BN_Leaky 是使用二維卷積定義的塊，可以去代碼里查看，很好理解的，

yolov3

yolov3使用了殘差結構和FPN, 網路較深,結構復雜，我們先來看一下他的整體網路結構：

這里使用的FPN(Feature Pyramid Network) 特征金字塔如下圖所示：

在目標檢測中，往往會包含不用大小的目標，多層卷積后，小目標的語意丟失比較嚴重，使用FPN 能有效的利用多層特征資訊 加強淺層小目標的特征新資訊，提升網路的檢測能力，

def yolo_body(inputs, num_anchors, num_classes):
    """Create YOLO_V3 model CNN body in Keras."""
    #------------------------------------------------------------------
	# inputs 輸入向量  num_anchors anchor boxes的數量 num_classes 類別數
	#------------------------------------------------------------------
    darknet = Model(inputs, darknet_body(inputs))
    x, y1 = make_last_layers(darknet.output, 512, num_anchors*(num_classes+5))

    x = compose(
            DarknetConv2D_BN_Leaky(256, (1,1)),
            UpSampling2D(2))(x)
    x = Concatenate()([x,darknet.layers[152].output])
    x, y2 = make_last_layers(x, 256, num_anchors*(num_classes+5))

    x = compose(
            DarknetConv2D_BN_Leaky(128, (1,1)),
            UpSampling2D(2))(x)
    x = Concatenate()([x,darknet.layers[92].output])
    x, y3 = make_last_layers(x, 128, num_anchors*(num_classes+5))
	# yolo head的輸出，
    return Model(inputs, [y1,y2,y3])

代碼中的 make_last_layers 是產生YOLO的輸出層，對于引數 ：

 num_anchors*(num_classes+5)

yolo3 每個網格點有3個anchor boxes，num_achors=3 ，每個anchor box都要預測所有的類別，假設我們使用的是Voc2007 有20類別，num_classes=20, 5代表框的p（框中有目標的概率），x_offset、y_offset、h和w 4個值，yolo head 中的輸出維度就為 [batch_size,w,h,3x(20+5)]，
到這里YOLO的主干網路解介紹完了，

二、配置訓練引數

1.目標檢測資料集

以VOC 2007 資料集為例,首先來看一下檔案樹：

─VOC2007
    ├─Annotations
    │	└─000005.xml
    │	└─000006.xml
    │	└─xxxx.xml	
    ├─ImageSets
    │  └─Main
    └─JPEGImages
    │	└─000005.jpg
    │	└─000006.jpg
    │	└─xxxx.jpg

每一個xml 包含同名jpg 中目標的位置以及類別，在訓練時，首先要將資料 通過 voc_annotation.py 轉換到記事本中，方便訓練時讀取，記事本形式如下：

VOC2007/JPEGImages/000005.jpg（圖片路徑）98,267,194,383（框的位置） ,1（框中目標的類別）
VOC2007/JPEGImages/000006.jpg（圖片路徑）99,205,198,318（框的位置） ,1（框中目標的類別）

這樣的資料是無法直接和YOLO的輸出進行計算的，那么我們還需要將，這種資料編碼成和yolo head 的格式一樣，才能去計算loss，反向傳播調整引數
通過如下函式將y_ture，轉換成和y_predict 一樣的形式：

def preprocess_true_boxes(true_boxes, input_shape, anchors, num_classes):
    '''Preprocess true boxes to training input format
    Parameters
    ----------
    	true_boxes: array, shape=(m, T, 5)
        Absolute x_min, y_min, x_max, y_max, class_id relative to input_shape.
    input_shape: array-like, hw, multiples of 32
    anchors: array, shape=(N, 2), wh
    num_classes: integer
    Returns
    -------
    y_true: list of array, shape like yolo_outputs, xywh are reletive value

    '''
    assert (true_boxes[..., 4]<num_classes).all(), 'class id must be less than num_classes'
    num_layers = len(anchors)//3 # default setting
    anchor_mask = [[6,7,8], [3,4,5], [0,1,2]] if num_layers==3 else [[3,4,5], [1,2,3]]

    true_boxes = np.array(true_boxes, dtype='float32')
    input_shape = np.array(input_shape, dtype='int32')
    boxes_xy = (true_boxes[..., 0:2] + true_boxes[..., 2:4]) // 2
    boxes_wh = true_boxes[..., 2:4] - true_boxes[..., 0:2]
    true_boxes[..., 0:2] = boxes_xy/input_shape[::-1]
    true_boxes[..., 2:4] = boxes_wh/input_shape[::-1]

    m = true_boxes.shape[0]
    grid_shapes = [input_shape//{0:32, 1:16, 2:8}[l] for l in range(num_layers)]
    y_true = [np.zeros((m,grid_shapes[l][0],grid_shapes[l][1],len(anchor_mask[l]),5+num_classes),
        dtype='float32') for l in range(num_layers)]

    # Expand dim to apply broadcasting.
    anchors = np.expand_dims(anchors, 0)
    anchor_maxes = anchors / 2.
    anchor_mins = -anchor_maxes
    valid_mask = boxes_wh[..., 0]>0

    for b in range(m):
        # Discard zero rows.
        wh = boxes_wh[b, valid_mask[b]]
        if len(wh)==0: continue
        # Expand dim to apply broadcasting.
        wh = np.expand_dims(wh, -2)
        box_maxes = wh / 2.
        box_mins = -box_maxes

        intersect_mins = np.maximum(box_mins, anchor_mins)
        intersect_maxes = np.minimum(box_maxes, anchor_maxes)
        intersect_wh = np.maximum(intersect_maxes - intersect_mins, 0.)
        intersect_area = intersect_wh[..., 0] * intersect_wh[..., 1]
        box_area = wh[..., 0] * wh[..., 1]
        anchor_area = anchors[..., 0] * anchors[..., 1]
        iou = intersect_area / (box_area + anchor_area - intersect_area)

        # Find best anchor for each true box
        best_anchor = np.argmax(iou, axis=-1)

        for t, n in enumerate(best_anchor):
            for l in range(num_layers):
                if n in anchor_mask[l]:
                    i = np.floor(true_boxes[b,t,0]*grid_shapes[l][1]).astype('int32')
                    j = np.floor(true_boxes[b,t,1]*grid_shapes[l][0]).astype('int32')
                    k = anchor_mask[l].index(n)
                    c = true_boxes[b,t, 4].astype('int32')
                    y_true[l][b, j, i, k, 0:4] = true_boxes[b,t, 0:4]
                    y_true[l][b, j, i, k, 4] = 1
                    y_true[l][b, j, i, k, 5+c] = 1
    #print(y_true.shape)
    return y_true

資料集放置如下圖所示：

2.設定anchor box 和classes

Yolo 中的anchor boxes 是通過資料集中框的大小通過kmeans聚類而來，yolov3 有三個輸出，每個網格預測3個，所以 Kmeans中的K設定9，如果是tiny版本的話，就設定為6，代碼如下：

import numpy as np
"""
Kmeans 聚類演算法， 根據 資料集中的xml 檔案 聚類出是和目標的anchor boxes
"""
class YOLO_Kmeans:

    def __init__(self, cluster_number, filename):
        self.cluster_number = cluster_number
        self.filename =filename

    def iou(self, boxes, clusters):  # 1 box -> k clusters
        n = boxes.shape[0]
        k = self.cluster_number

        box_area = boxes[:, 0] * boxes[:, 1]
        box_area = box_area.repeat(k)
        box_area = np.reshape(box_area, (n, k))

        cluster_area = clusters[:, 0] * clusters[:, 1]
        cluster_area = np.tile(cluster_area, [1, n])
        cluster_area = np.reshape(cluster_area, (n, k))

        box_w_matrix = np.reshape(boxes[:, 0].repeat(k), (n, k))
        cluster_w_matrix = np.reshape(np.tile(clusters[:, 0], (1, n)), (n, k))
        min_w_matrix = np.minimum(cluster_w_matrix, box_w_matrix)

        box_h_matrix = np.reshape(boxes[:, 1].repeat(k), (n, k))
        cluster_h_matrix = np.reshape(np.tile(clusters[:, 1], (1, n)), (n, k))
        min_h_matrix = np.minimum(cluster_h_matrix, box_h_matrix)
        inter_area = np.multiply(min_w_matrix, min_h_matrix)

        result = inter_area / (box_area + cluster_area - inter_area)
        return result

    def avg_iou(self, boxes, clusters):
        accuracy = np.mean([np.max(self.iou(boxes, clusters), axis=1)])
        return accuracy

    def kmeans(self, boxes, k, dist=np.median):
        box_number = boxes.shape[0]
        distances = np.empty((box_number, k))
        last_nearest = np.zeros((box_number,))
        np.random.seed()
        clusters = boxes[np.random.choice(
            box_number, k, replace=False)]  # init k clusters
        while True:

            distances = 1 - self.iou(boxes, clusters)

            current_nearest = np.argmin(distances, axis=1)
            if (last_nearest == current_nearest).all():
                break  # clusters won't change
            for cluster in range(k):
                clusters[cluster] = dist(  # update clusters
                    boxes[current_nearest == cluster], axis=0)

            last_nearest = current_nearest

        return clusters

    def result2txt(self, data):
        f = open("yolo_anchors1.txt", 'w') # 這里是生存achor后保存的路徑
        row = np.shape(data)[0]
        for i in range(row):
            if i == 0:
                x_y = "%d,%d" % (data[i][0], data[i][1])
            else:
                x_y = ", %d,%d" % (data[i][0], data[i][1])
            f.write(x_y)
        f.close()

    def txt2boxes(self):
        f = open(self.filename, 'r')
        dataSet = []
        for line in f:
            infos = line.split(" ")
            length = len(infos)
            for i in range(1, length):
                width = int(infos[i].split(",")[2]) - \
                    int(infos[i].split(",")[0])
                height = int(infos[i].split(",")[3]) - \
                    int(infos[i].split(",")[1])
                dataSet.append([width, height])
        result = np.array(dataSet)
        f.close()
        return result

    def txt2clusters(self):
        all_boxes = self.txt2boxes()
        result = self.kmeans(all_boxes, k=self.cluster_number)
        result = result[np.lexsort(result.T[0, None])]
        self.result2txt(result)
        print("K anchors:\n {}".format(result))
        print("Accuracy: {:.2f}%".format(
            self.avg_iou(all_boxes, result) * 100))

#通過聚類方法設定資料集中合適的框
if __name__ == "__main__":
    cluster_number =9 #聚類框的個數 6 或者 9
    filename = r"2007_train.txt" # 指定train.txt的路勁
    kmeans = YOLO_Kmeans(cluster_number, filename)
    kmeans.txt2clusters()

當然也可以使用yolo中默認的achor box的大小，
classes.txt 是宣告 你目標中所有的類別，以戴口罩資料集為例

without_mask
with_mask
mask_weared_incorrect

文本中書寫的類別要和你xml 檔案中所寫類別保持一致，才能正確索引，

三、 配置訓練程序

訓練程序需要配置的東西，就是之前的資料準備，輸入模型即可，
訓練過錯記憶體溢位，記得吧batch_size 改小一點，yolo完整版的話一般設定為8比較好
代碼：

def train():
    # 你的 資料集檔案路勁
    train_annotation_path = r"2007_train.txt" # 生存的資料索引txt文本
    val_annotation_path = r"2007_val.txt"

    anchors_path = r"model_data\mask_anchor.txt" # 生成的anchor
    classes_path = r"model_data\mask_classes.txt" #自己資料集的類別
    log_dir = "logs/tiny_log/"
    weights_dir = "weights/"
    class_names = get_classes(classes_path)
    num_classes = len(class_names)
    anchors = get_anchors(anchors_path)

    input_shape = (416, 416)  #  必須是32的 倍數  yolo 的設定

	# tiny 版本
    model = create_tiny_model(
        input_shape=input_shape,
        anchors=anchors,
        num_classes=num_classes,
        freeze_body=2,
        weights_path="model_data\yolov3-tiny.h5",
    )
    ''''
     # yolo v3 的完整版本 ，
    model = create_model(
        input_shape=input_shape,
        anchors=anchors,
        num_classes=num_classes,
        # freeze_body=0,
        weights_path="model_data\yolov3.h5",
    )
    '''
    print(len(model.layers))
    logging = TensorBoard(log_dir=log_dir)
    checkpoint = ModelCheckpoint(
        weights_dir + "ep{epoch:03d}-loss{loss:.3f}-val_loss{val_loss:.3f}.h5",
        monitor="val_loss",
        save_weights_only=True,
        save_best_only=True,
        period=3,
    )
    reduce_lr = ReduceLROnPlateau(monitor="val_loss", factor=0.1, patience=3, verbose=1)
    early_stopping = EarlyStopping(
        monitor="val_loss", min_delta=0, patience=10, verbose=1
    )

    # 讀取資料集對應的 .txt 檔案
    with open(train_annotation_path) as f:
        train_lines = f.readlines()
    with open(val_annotation_path) as f:
        val_lines = f.readlines()
    num_train = len(train_lines)
    num_val = len(val_lines)

    # 配置訓練引數
    Freeze_Train = True
    # ------------------------------------------------------------
    # 先凍結一定網路層進行訓練，這樣訓練比較快 ,得到一個loss穩定的model
    # -------------------------------------------------------------
    if Freeze_Train:
        batch_size = 32
        model.compile(
            optimizer=Adam(1e-3),
            loss={# use custom yolo_loss Lambda layer.
            "yolo_loss": lambda y_true, y_pred: y_pred}
        )
        print("Train on {} samples, val on {} samples, with batch size {}.".format(num_train, num_val, batch_size))
        model.fit(
            data_generator_wrapper(train_lines,batch_size,input_shape,anchors,num_classes),
            steps_per_epoch=max(1,num_train//batch_size),
            validation_data=data_generator_wrapper(val_lines,batch_size,input_shape,anchors,num_classes),
            validation_steps=max(1,num_val//batch_size),
            epochs=50,
            initial_epoch=0,
            callbacks=[logging,checkpoint]
        )
        model.save_weights(weights_dir+'trained_weights_stage_1.h5')
     #-----------------------------------------------------------------
     # 解凍所有層，并調小學習率訓練
     #-----------------------------------------------------------------
    if True:
        for i in range(len(model.layers)): model.layers[i].trainable = True
        model.compile(optimizer=Adam(lr=1e-4), loss={'yolo_loss': lambda y_true, y_pred: y_pred}) # recompile to apply the change
        print('Unfreeze all of the layers.')

        batch_size = 32 # note that more GPU memory is required after unfreezing the body
        print('Train on {} samples, val on {} samples, with batch size {}.'.format(num_train, num_val, batch_size))
        model.fit_generator(data_generator_wrapper(train_lines, batch_size, input_shape, anchors, num_classes),
            steps_per_epoch=max(1, num_train//batch_size),
            validation_data=data_generator_wrapper(val_lines, batch_size, input_shape, anchors, num_classes),
            validation_steps=max(1, num_val//batch_size),
            epochs=100,
            initial_epoch=50,
            callbacks=[logging, checkpoint, reduce_lr, early_stopping])
        model.save_weights(log_dir + 'trained_weights_final.h5')


if __name__=='__main__':
    train()

上述就是訓練程序了，訓練時，
注意：
使用model 時，tiny 和YOLOv3 achor 不一樣，一個是6 類 一個是9類 ，記得更換，
tips:訓練過錯先凍結一部分層去訓練，這樣訓練比較快，當損失穩定之后，使用更小的學習率去，趨勢模型收斂更好，
訓練結果如下：

可以看出收斂效果還是相當不錯的，，，，

四、模型預測

模型預測，將圖片輸入到保存的模型當中，如果輸出一個跟yolo head 一樣的維度，很顯然這個時候我們是無法獲圖片的類別還有框的，那么需要通過解碼，拿到我們有用的資料，這個作業在專案中的yolo.py 中實作，

#這里不給出全部代碼了，但在預測是，網路的權重檔案 achor boxes 和classes 要保持一直，在檔案中這個位置配置
_defaults = {
        # ----------------------------------
        # 模型路徑, anchor路徑，class 路徑， 
        # sorce 一本設為0.5  置信度閾值 iou 
        #  輸入圖片大小 一般是13 的倍數
        #   gpu 1
        #  遇上維度 不匹配問題  可能是訓練時 忘記讓model_path 和class_path 對應 
        # -----------------------------------
        "model_path": r"logs\yolov3_log\trained_weights_final.h5",
        "anchors_path": "model_data\mask_anchor.txt",
        "classes_path": "model_data\mask_classes.txt",
        "score": 0.3,
        "iou": 0.3,
        "model_image_size": (416, 416),
        "gpu_num": 1, # 這里使用1 ，多gpu 由于TensorFlow 版本問題 我給注解掉了
    }

模型預測完整代碼：

import time

import cv2
import numpy as np
import tensorflow as tf
from PIL import Image
from yolo import YOLO,detect_video
import os
from tqdm import tqdm
def predict():
    #-------------------------------------------------------------
    # 新建 yolo 物件，使用模型權重路徑，以及其他引數，到yolo.py 中更改
    # predict_model 為 img 預測圖片， video  輸入為視頻，記得更改視頻路勁   video_path 為 0 時呼叫攝像頭  為視頻路徑時 讀取視頻
    # predict_model 為 dir_predict 時輸入為存放圖片的路徑 ，預測完成后 放入out_img 使用時記得修改路勁
    # 要預測  指定類別是 ，可以設定好重新訓練模型，或者進入detect_image 修改引數 if predicted_classes='car'
    #--------------------------------------------------------------
    yolo=YOLO()  
    predict_model='dir_predict'

    video_path= 0
    video_save_path=""

    dir_img_input='img/'
    dir_save_path='out_img/'
    if predict_model=='img':
        while(True):
            img=input('Input image filename:')
            try:
                image=Image.open(img)
            except:
                print('Open Image Error！ Please Try Again')
                continue
            else:
                out_image=yolo.detect_image(image)
                out_image.show()
    elif predict_model=='video':
        detect_video(yolo,video_path,video_save_path) # 可以獲取fps
    
    elif predict_model=='dir_predict':
        #---------------------------------
        #拿到所有圖片 通過detect image 檢測
        #-------------------------------
        imgs=os.listdir(dir_img_input)
        for img_name in tqdm(imgs):
             if img_name.lower().endswith(('.bmp', '.dib', '.png', '.jpg', '.jpeg', '.pbm', '.pgm', '.ppm', '.tif', '.tiff')):
                image_path  = os.path.join(dir_img_input, img_name)
                image       = Image.open(image_path)
                r_image     = yolo.detect_image(image)
                if not os.path.exists(dir_save_path):
                    os.makedirs(dir_save_path)
                r_image.save(os.path.join(dir_save_path, img_name)) 
    else:
           raise AssertionError("Please specify the correct mode: 'img', 'video', 'dir_predict'.")
if __name__=='__main__':
    predict()

使用方法寫在注釋里了，支持視頻，圖片 檔案夾
檢測結果：

總結

目標檢測網路結構，其實跟影像分類差不多，但對資料讀取程序，編碼解碼程序比較復雜，對于這個程序，我掌握不是很多，所以就不在這里講解了，
目標檢測訓練比較難，資料標注比較麻煩，后面可能會出一期視頻，教大家如何配置，所有代碼和資源均會上傳，寫代碼不易，歡迎點贊~

博主的抱怨

哎，太難了，哎~ cv 太卷了，