CVPR2021配准算法LoFTR的配置（LoFTR: Detector-Free Local Feature Matching with Transformers）

最新推荐文章于 2024-08-21 15:48:33 发布

博博有个大大大的Dream

最新推荐文章于 2024-08-21 15:48:33 发布

阅读量9.5k

点赞数 13

分类专栏：图像配准文章标签：算法深度学习机器学习

本文链接：https://blog.csdn.net/qq_17783559/article/details/120599526

版权

图像配准专栏收录该内容

1 篇文章 2 订阅

订阅专栏

这篇博客介绍了如何下载和运行LoFTR，一个用于视觉特征匹配的深度学习模型。首先，从给定链接下载论文和代码，然后创建并激活Python虚拟环境，安装必要的库。接着，下载额外的superglue.py文件和预训练模型。在配置好环境后，通过运行mydemo.py脚本，应用LoFTR对示例图片进行匹配，并生成匹配结果的PDF文件。整个过程详细展示了LoFTR在实际应用中的步骤。

摘要由CSDN通过智能技术生成

1、论文下载地址：

https://arxiv.org/pdf/2104.00680.pdf

2、代码下载地址：

https://github.com/zju3dv/LoFTR

3、新建虚拟python环境并激活

conda create -n LoFTR python=3.7
source activate LoFTR

4、安装需要的库

pip install torch==1.6.0 einops yacs kornia opencv-python matplotlib

5、下载额外库

在如下网址https://raw.githubusercontent.com/magicleap/SuperGluePretrainedNetwork/master/models/superglue.py

下载superglue.py

然后放到路径src/loftr/utils

6、下载预训练模型

链接：https://pan.baidu.com/s/1dwUDx6A9lRMBkCSowLIz5Q
提取码：jlcl
放到路径：weights/outdoor_ds.ckpt

7、demo运行

1）加入工程路径环境变量：

export PYTHONPATH=$PYTHONPATH:/home1/users/XXX/Codes/LoFTR-master/

2）新建mydemo.py写入如下代码：

import os
os.chdir("..")
from copy import deepcopy

# import os
# os.environ["CUDA_VISIBLE_DEVICES"] = "0"

import torch
import cv2
import numpy as np
import matplotlib.cm as cm
from src.utils.plotting import make_matching_figure


from src.loftr import LoFTR, default_cfg

# The default config uses dual-softmax.
# The outdoor and indoor models share the same config.
# You can change the default values like thr and coarse_match_type.

matcher = LoFTR(config=default_cfg)
matcher.load_state_dict(torch.load("/home1/users/XXX/Codes/LoFTR-master/weights/outdoor_ds.ckpt")['state_dict'])
matcher = matcher.eval().cuda()


default_cfg['coarse']

# Load example images
img0_pth = "/home1/users/XXX/Codes/LoFTR-master/assets/phototourism_sample_images/united_states_capitol_26757027_6717084061.jpg"
img1_pth = "/home1/users/XXX/Codes/LoFTR-master/assets/phototourism_sample_images/united_states_capitol_98169888_3347710852.jpg"
img0_raw = cv2.imread(img0_pth, cv2.IMREAD_GRAYSCALE)
img1_raw = cv2.imread(img1_pth, cv2.IMREAD_GRAYSCALE)
img0_raw = cv2.resize(img0_raw, (img0_raw.shape[1]//8*8, img0_raw.shape[0]//8*8))  # input size shuold be divisible by 8
img1_raw = cv2.resize(img1_raw, (img1_raw.shape[1]//8*8, img1_raw.shape[0]//8*8))

img0 = torch.from_numpy(img0_raw)[None][None].cuda() / 255.
img1 = torch.from_numpy(img1_raw)[None][None].cuda() / 255.
batch = {'image0': img0, 'image1': img1}

# Inference with LoFTR and get prediction
with torch.no_grad():
    matcher(batch)
    mkpts0 = batch['mkpts0_f'].cpu().numpy()
    mkpts1 = batch['mkpts1_f'].cpu().numpy()
    mconf = batch['mconf'].cpu().numpy()

# Draw
color = cm.jet(mconf)
text = [
    'LoFTR',
    'Matches: {}'.format(len(mkpts0)),
]
fig = make_matching_figure(img0_raw, img1_raw, mkpts0, mkpts1, color, text=text,path="LoFTR-master/LoFTR-colab-demo.pdf")

3）运行python mydemo.py得到结果：