DXSLAM是一个基于深度CNN特征提取的视觉SLAM系统,论文地址,实测效果不咋地,TUM换个数据集就跟踪丢了
C++11 or C++0x Compiler
Pangolin
OpenCV
Eigen3
Dbow、Fbow and g2o (Included in Thirdparty folder)
tensorflow(1.12)
git clone https://github.com/raulmur/DXSLAM.git DXSLAM
cd dxslam
chmod +x build.sh
./build.sh
TUM数据集下载地址,其中前缀fr1的是TUM1数据集、前缀fr2的是TUM2数据集、前缀fr3的是TUM3数据集,这里选择下载fr1/xyz
(很多数据集都跑丢失,这个数据集可以跑成功)
下载好数据集后,解压到DXSLAM
项目中,需要同步时间戳:
在解压后的数据集文件夹内创建一个 associate.py
脚本用于同步时间戳,基于Python2写的:
#!/usr/bin/python
# Software License Agreement (BSD License)
#
# Copyright (c) 2013, Juergen Sturm, TUM
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following
# disclaimer in the documentation and/or other materials provided
# with the distribution.
# * Neither the name of TUM nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
# FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
# COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
# BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
# ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
# POSSIBILITY OF SUCH DAMAGE.
#
# Requirements:
# sudo apt-get install python-argparse
"""
The Kinect provides the color and depth images in an un-synchronized way. This means that the set of time stamps from the color images do not intersect with those of the depth images. Therefore, we need some way of associating color images to depth images.
For this purpose, you can use the ''associate.py'' script. It reads the time stamps from the rgb.txt file and the depth.txt file, and joins them by finding the best matches.
"""
import argparse
import sys
import os
import numpy
def read_file_list(filename):
"""
Reads a trajectory from a text file.
File format:
The file format is "stamp d1 d2 d3 ...", where stamp denotes the time stamp (to be matched)
and "d1 d2 d3.." is arbitary data (e.g., a 3D position and 3D orientation) associated to this timestamp.
Input:
filename -- File name
Output:
dict -- dictionary of (stamp,data) tuples
"""
file = open(filename)
data = file.read()
lines = data.replace(","," ").replace("\t"," ").split("\n")
list = [[v.strip() for v in line.split(" ") if v.strip()!=""] for line in lines if len(line)>0 and line[0]!="#"]
list = [(float(l[0]),l[1:]) for l in list if len(l)>1]
return dict(list)
def associate(first_list, second_list,offset,max_difference):
"""
Associate two dictionaries of (stamp,data). As the time stamps never match exactly, we aim
to find the closest match for every input tuple.
Input:
first_list -- first dictionary of (stamp,data) tuples
second_list -- second dictionary of (stamp,data) tuples
offset -- time offset between both dictionaries (e.g., to model the delay between the sensors)
max_difference -- search radius for candidate generation
Output:
matches -- list of matched tuples ((stamp1,data1),(stamp2,data2))
"""
first_keys = first_list.keys()
second_keys = second_list.keys()
potential_matches = [(abs(a - (b + offset)), a, b)
for a in first_keys
for b in second_keys
if abs(a - (b + offset)) < max_difference]
potential_matches.sort()
matches = []
for diff, a, b in potential_matches:
if a in first_keys and b in second_keys:
first_keys.remove(a)
second_keys.remove(b)
matches.append((a, b))
matches.sort()
return matches
if __name__ == '__main__':
# parse command line
parser = argparse.ArgumentParser(description='''
This script takes two data files with timestamps and associates them
''')
parser.add_argument('first_file', help='first text file (format: timestamp data)')
parser.add_argument('second_file', help='second text file (format: timestamp data)')
parser.add_argument('--first_only', help='only output associated lines from first file', action='store_true')
parser.add_argument('--offset', help='time offset added to the timestamps of the second file (default: 0.0)',default=0.0)
parser.add_argument('--max_difference', help='maximally allowed time difference for matching entries (default: 0.02)',default=0.02)
args = parser.parse_args()
first_list = read_file_list(args.first_file)
second_list = read_file_list(args.second_file)
matches = associate(first_list, second_list,float(args.offset),float(args.max_difference))
if args.first_only:
for a,b in matches:
print("%f %s"%(a," ".join(first_list[a])))
else:
for a,b in matches:
print("%f %s %f %s"%(a," ".join(first_list[a]),b-float(args.offset)," ".join(second_list[b])))
运行脚本:
python associate.py PATH_TO_SEQUENCE/rgb.txt PATH_TO_SEQUENCE/depth.txt > associations.txt
# 例如
cd rgbd_dataset_freiburg1_xyz
python associate.py rgb.txt depth.txt > associations.txt # 需要Python2,python3会报错
TUM
的配置文件因为DXSLAM
是基于ORB2
改的,这里直接复制ORB2
里面的TUM1
的配置文件即可(TUM1
配置文件对应哪些fr1
前缀的数据集,如果下的是其他系列的数据集就复制相应的配置文件)
这里给出TUM1
配置文件内容:
%YAML:1.0
#--------------------------------------------------------------------------------------------
# Camera Parameters. Adjust them!
#--------------------------------------------------------------------------------------------
File.version: "1.0"
Camera.type: "PinHole"
# Camera calibration and distortion parameters (OpenCV)
Camera.fx: 517.306408
Camera.fy: 516.469215
Camera.cx: 318.643040
Camera.cy: 255.313989
Camera.k1: 0.262383
Camera.k2: -0.953104
Camera.p1: -0.005358
Camera.p2: 0.002628
Camera.k3: 1.163314
Camera.width: 640
Camera.height: 480
# Camera frames per second
Camera.fps: 30
# Color order of the images (0: BGR, 1: RGB. It is ignored if images are grayscale)
Camera.RGB: 1
# Close/Far threshold. Baseline times.
Stereo.ThDepth: 40.0
Stereo.b: 0.07732
# Depth map values factor
RGBD.DepthMapFactor: 5000.0 # 1.0 for ROS_bag
#--------------------------------------------------------------------------------------------
# ORB Parameters
#--------------------------------------------------------------------------------------------
# ORB Extractor: Number of features per image
ORBextractor.nFeatures: 1000
# ORB Extractor: Scale factor between levels in the scale pyramid
ORBextractor.scaleFactor: 1.2
# ORB Extractor: Number of levels in the scale pyramid
ORBextractor.nLevels: 8
# ORB Extractor: Fast threshold
# Image is divided in a grid. At each cell FAST are extracted imposing a minimum response.
# Firstly we impose iniThFAST. If no corners are detected we impose a lower value minThFAST
# You can lower these values if your images have low contrast
ORBextractor.iniThFAST: 20
ORBextractor.minThFAST: 7
#--------------------------------------------------------------------------------------------
# Viewer Parameters
#--------------------------------------------------------------------------------------------
Viewer.KeyFrameSize: 0.05
Viewer.KeyFrameLineWidth: 1.0
Viewer.GraphLineWidth: 0.9
Viewer.PointSize: 2.0
Viewer.CameraSize: 0.08
Viewer.CameraLineWidth: 3.0
Viewer.ViewpointX: 0.0
Viewer.ViewpointY: -0.7
Viewer.ViewpointZ: -1.8
Viewer.ViewpointF: 500.0
该项目是离线运行的,需要提前对数据集提取特征点,深度学习环境需要TensorFlow1.12,这里我安装1.14也可以通过:
# 创建一个conda虚拟环境,安装TensorFlow1.14
conda create -n tensorflow114 python=3.7
pip install tensorflow_gpu==1.14.0
pip install keras==2.2.5
pip install protobuf==3.20.0
pip install numpy==1.16.5
pip install pandas==1.0.0
pip install sklearn
pip install matplotlib==3.0.0
pip install python-opencv
配置好环境,就可以运行深度学习模型,提取特征点了:
cd hf-net
python3 getFeature.py image/path/to/rgb output/feature/path
# 例如
python getFeature.py ../rgbd_dataset_freiburg1_xyz/rgb ../feature
先看一下此时的项目文件组成:(只显示主要文件)
├── Examples
│ └── RGB-D
│ ├── rgbd_tum (主程序)
│ ├── rgbd_tum.cc
│ ├── TUM1.yaml (配置文件)
├── feature (深度学习模型提取的特征点)
├── rgbd_dataset_freiburg1_xyz (TUM数据集)
└── Vocabulary
├── DXSLAM.fbow (词袋模型)
└── DXSLAM.tar.xz
./Examples/RGB-D/rgbd_tum Vocabulary/DXSLAM.fbow Examples/RGB-D/TUMX.yaml PATH_TO_SEQUENCE_FOLDER ASSOCIATIONS_FILE OUTPUT/FEATURE/PATH
# 例如
./Examples/RGB-D/rgbd_tum Vocabulary/DXSLAM.fbow ./Examples/RGB-D/TUM1.yaml ./rgbd_dataset_freiburg1_xyz ./rgbd_dataset_freiburg1_xyz/associations.txt ./feature