open-mmlab的检测项目
数据标柱工具介绍
labelImg
git地址: https://github.com/tzutalin/labelImg
安装(两种方法都适用于linux和mac上的conda虚拟环境下安装)
根据git上介绍,有两种安装方法:
pip包安装
pip包安装(linux上最简单的安装方法),安装命令及使用命令如下:
# 环境准备
conda create -n labelImg_1 python=3.7
conda activate labelImg_1
# linux安装
pip3 install labelImg
# linux使用vnc打开可视化界面。命令行输入如下:
labelImg
#
labelImg [IMAGE_PATH] [PRE-DEFINED CLASS FILE]
使用pip list查看安装包
pip list
Package Version
---------- ---------
certifi 2021.10.8
labelImg 1.8.6
lxml 4.6.3
pip 21.0.1
PyQt5 5.15.5
PyQt5-Qt5 5.15.2
PyQt5-sip 12.9.0
setuptools 58.0.4
wheel 0.37.0
软件截图如下
从源码安装
下载源代码,准备虚拟环境
# 环境准备
conda create -n labelImg_2 python==3.7
conda activate labelImg_2
pip install PyQt5
pip install lxml
pyrcc5 -o libs/resources.py resources.qrc
# 下载
git clone https://github.com/tzutalin/labelImg
cd labelImg
# 使用
python3 labelImg.py
#
python3 labelImg.py [IMAGE_PATH] [PRE-DEFINED CLASS FILE]
使用pip list查看安装包
pip list
Package Version
---------- ---------
certifi 2021.10.8
lxml 4.6.3
pip 21.2.2
PyQt5 5.15.5
PyQt5-Qt5 5.15.2
PyQt5-sip 12.9.0
setuptools 58.0.4
wheel 0.37.0
软件截图如下
labelme
git地址: https://github.com/wkentaro/labelme
安装
安装命令如下:
conda create --name=labelme python=3.6
source activate labelme
pip install labelme
使用
直接在环境下键入
labelme
出现界面如下,此时可以打开图片路径进行标注:
点击Create Polygons进行标注,输入标签
点击左侧菜单栏保存,保存为json文件。
生成label数据
# 运用命令
labelme_json_to_dataset DJI_20210414132518_0423_Z.json
生成文件夹DJI_20210414132518_0423_Z_json,内容如下:
注:可以把多个labelme_json_to_dataset aa.json写入批处理文件中
mmdetection
git地址: https://github.com/open-mmlab/mmdetection
git clone https://github.com/open-mmlab/mmdetection
下载到服务器,更新到本地,用pycharm编辑器打开(参考https://blog.csdn.net/fighting_Kitty/article/details/121023315)
配置环境
找到readme中Installation,get_started.md中详细写了环境安装,主要命令如下:
# 创建虚拟环境
conda create -n openmmlab python=3.7 -y
conda activate openmmlab
# 安装PyTorch and torchvision
#(按照自己的cuda版本(nvcc -V)
# 去pytorch官网上选https://pytorch.org/get-started/previous-versions/)
# CUDA 11.0
pip install torch==1.7.1+cu110 torchvision==0.8.2+cu110 torchaudio==0.7.2 -f https://download.pytorch.org/whl/torch_stable.html
# Install MMDetection(自动或者手动,自动简单,手动适合调试研究代码)
pip install openmim
mim install mmdet
# 手动
pip install torch==1.7.0+cu110 torchvision==0.8.1+cu110 torchaudio==0.7.0 -f https://download.pytorch.org/whl/torch_stable.html
pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu110/1.7.1+cu110/index.htm
附:cuda查看nvcc -V
nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2020 NVIDIA Corporation
Built on Thu_Jun_11_22:26:38_PDT_2020
Cuda compilation tools, release 11.0, V11.0.194
Build cuda_11.0_bu.TC445_37.28540450_0
运行
mkdir checkpoints
cd checkpoints
wget https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth
wget https://github.com/open-mmlab/mmdetection/tree/master/configs/faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py
代码
from mmdet.apis import init_detector, inference_detector
config_file = '../configs/faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py'
# download the checkpoint from model zoo and put it in `checkpoints/`
# url: https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth
checkpoint_file = '../checkpoints/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth'
device = 'cuda:0'
# init a detector
model = init_detector(config_file, checkpoint_file, device=device)
# inference the demo image
inference_detector(model, '../demo/demo.jpg')
用自己的数据集微调
下载预训练模型
mmdetection里提供了丰富的预训练权重模型,本文使用faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth,下载后放到checkpoints路径下
voc数据修改为标准的coco格式
用labelImg标注检测图片,标注数据格式为PASCAL VOC的格式,一个图片对应一个XML注文件。COCO数据格式,所有的图像数据标注信息都保存在一个JSON文件中。转换代码如下:
import json
import os, sys
import xml.etree.ElementTree as ET
import pdb;pdb.set_trace()
mode = 'test' # 这里可以改成train、val
DATA_DIR = '/home/qian/caojie/software/mmdetection/data/hotel/' + mode + '/'
ANN_DIR = DATA_DIR + 'annotations/'
# ==================== 需要修改 train or val ========================
COCO_JSON_FILE = DATA_DIR + 'hotel_' + mode + '.json' # json save path
VOC_XMLS_DIR = DATA_DIR + 'annotations/'
# ==================================================================
if not os.path.exists(ANN_DIR):
os.makedirs(ANN_DIR)
# coco images 的列表
images = []
# coco annotations 的列表
annotations = []
# coco categories 的列表
# If necessary, pre-define category and its id
PRE_DEFINE_CATEGORIES = {
"background": 0, "bed": 1, "chair": 2, "curtain": 3}
categories = [
{
'id': 0,
'name':