目标检测 label_目标检测——制作GroundTruth！开始训练！

本专栏目录：

核心内容篇：

• 目标检测——从简单的开始！进击的YOLO！
• 目标检测——Backbone与Detection head
• 目标检测——搭建更好更快的YOLO！
• 目标检测——制作GroundTruth！开始训练！
• 目标检测——YOLOv2！
• 目标检测——YOLOv3！
• 目标检测——YOLO轻量化初尝试-SlimYOLOv2
• 目标检测——小小的TinyYOLOv3
• 目标检测——anchor-free的回归之作-FCOS
• 目标检测——anchor-free之CenterNet的浅析与实现

番外篇：

• 番外一：darknet19与darknet53
• 番外二：ImageNet预训练
• 番外三：人脸检测实战-YOLAF
• 番外四：实时-轻量人脸检测
• 番外五：YOLOv3-Plus
• 番外六：YOLOv3-Slim
• 番外七：更好的backbone-CSPDarknet系列

---

1. 制作训练数据

本工作中，让我们来使用PASCAL VOC2007和VOC2012数据集作为训练集，VOC2007test作为测试集，之所以选择VOC，主要是因为它太经典了！而且VOC2007测试集还有真实标签，方便我们算mAP去评价模型的好坏。关于数据集的下载，大家可以执行下面的VOC2007.sh和VOC2012.sh两个批处理文件来下载数据：

VOC2007.sh:

#!/bin/bash
# Ellis Brown

start=`date +%s`

# handle optional download dir
if [ -z "$1" ]
  then
    # navigate to ~/data
    echo "navigating to ~/data/ ..." 
    mkdir -p ~/data
    cd ~/data/
  else
    # check if is valid directory
    if [ ! -d $1 ]; then
        echo $1 "is not a valid directory"
        exit 0
    fi
    echo "navigating to" $1 "..."
    cd $1
fi

echo "Downloading VOC2007 trainval ..."
# Download the data.
curl -LO http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtrainval_06-Nov-2007.tar
echo "Downloading VOC2007 test data ..."
curl -LO http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtest_06-Nov-2007.tar
echo "Done downloading."

# Extract data
echo "Extracting trainval ..."
tar -xvf VOCtrainval_06-Nov-2007.tar
echo "Extracting test ..."
tar -xvf VOCtest_06-Nov-2007.tar
echo "removing tars ..."
rm VOCtrainval_06-Nov-2007.tar
rm VOCtest_06-Nov-2007.tar

end=`date +%s`
runtime=$((end-start))

echo "Completed in" $runtime "seconds"

VOC2012.sh:

#!/bin/bash
# Ellis Brown

start=`date +%s`

# handle optional download dir
if [ -z "$1" ]
  then
    # navigate to ~/data
    echo "navigating to ~/data/ ..." 
    mkdir -p ~/data
    cd ~/data/
  else
    # check if is valid directory
    if [ ! -d $1 ]; then
        echo $1 "is not a valid directory"
        exit 0
    fi
    echo "navigating to" $1 "..."
    cd $1
fi

echo "Downloading VOC2012 trainval ..."
# Download the data.
curl -LO http://host.robots.ox.ac.uk/pascal/VOC/voc2012/VOCtrainval_11-May-2012.tar
echo "Done downloading."


# Extract data
echo "Extracting trainval ..."
tar -xvf VOCtrainval_11-May-2012.tar
echo "removing tar ..."
rm VOCtrainval_11-May-2012.tar

end=`date +%s`
runtime=$((end-start))

echo "Completed in" $runtime "seconds"

当然，COCO的下载文件我也提供了，显卡多的小伙伴可以考虑一下COCO呦~

假定，我们现在已经下好了数据集，VOC数据集中，图片是jpg格式的，而label都存在了xml文件中。如果读者在哪里看到了类似“VOC格式的label”，其实指的就是用xml文件来保存label数据。为了能够从xml文件读取出我们想要的bbox、class label等信息，我们首先定义一个类：

from .config import HOME
import os.path as osp
import sys
import torch
import torch.utils.data as data
import cv2
import numpy as np
if sys.version_info[0] == 2:
    import xml.etree.cElementTree as ET
else:
    import xml.etree.ElementTree as ET

VOC_CLASSES = (  # always index 0
    'aeroplane', 'bicycle', 'bird', 'boat',
    'bottle', 'bus', 'car', 'cat', 'chair',
    'cow', 'diningtable', 'dog', 'horse',
    'motorbike', 'person', 'pottedplant',
    'sheep', 'sofa', 'train', 'tvmonitor')

# note: if you used our download scripts, this should be right
VOC_ROOT = "e:python_workODyolo-guidedataVOCdevkit"


class VOCAnnotationTransform(object):
    """Transforms a VOC annotation into a Tensor of bbox coords and label index
    Initilized with a dictionary lookup of classnames to indexes

    Arguments:
        class_to_ind (dict, optional): dictionary lookup of classnames -> indexes
            (default: alphabetic indexing of VOC's 20 classes)
        keep_difficult (bool, optional): keep difficult instances or not
            (default