## Introduction
This is the PyTorch implementation of **AtLoc**, a simple and efficient neural architecture for robust visual localization.
#### Demos and Qualitative Results (click below for the video)
<p align="center"> <a href="https://youtu.be/_8NQXBadklU"><img src="./figures/real.gif" width="100%"></a> </p>
## Setup
AtLoc uses a Conda environment that makes it easy to install all dependencies.
1. Install [miniconda](https://docs.conda.io/en/latest/miniconda.html) with Python 2.7.
2. Create the `AtLoc` Conda environment: `conda env create -f environment.yml`.
3. Activate the environment: `conda activate py27pt04`.
4. Note that our code has been tested with PyTorch v0.4.1 (the environment.yml file should take care of installing the appropriate version).
## Data
We support the [7Scenes](https://www.microsoft.com/en-us/research/project/rgb-d-dataset-7-scenes/) and [Oxford RobotCar](http://robotcar-dataset.robots.ox.ac.uk/) datasets right now. You can also write your own PyTorch dataloader for other datasets and put it in the `data` directory.
### Special instructions for RobotCar:
1. Download [this fork](https://github.com/samarth-robo/robotcar-dataset-sdk/tree/master) of the dataset SDK, and run `cd data && ./robotcar_symlinks.sh` after editing the `ROBOTCAR_SDK_ROOT` variable in it appropriately.
2. For each sequence, you need to download the `stereo_centre`, `vo` and `gps` tar files from the dataset website. The directory for each 'scene' (e.g. `loop`) has .txt files defining the train/test_split.
3. To make training faster, we pre-processed the images using `data/process_robotcar.py`. This script undistorts the images using the camera models provided by the dataset, and scales them such that the shortest side is 256 pixels.
4. Pixel and Pose statistics must be calculated before any training. Use the `data/dataset_mean.py`, which also saves the information at the proper location. We provide pre-computed values for RobotCar and 7Scenes.
## Running the code
### Training
The executable script is `train.py`. For example:
- AtLoc on `loop` from `RobotCar`:
```
python train.py --dataset RobotCar --scene loop --model AtLoc --gpus 0
```
- AtLocLstm on `loop` from `RobotCar`:
```
python train.py --dataset RobotCar --scene loop --model AtLoc --lstm True --gpus 0
```
- AtLoc+ on `loop` from `RobotCar`:
```
python train.py --dataset RobotCar --scene loop --model AtLocPlus --gamma -3.0 --gpus 0
```
The meanings of various command-line parameters are documented in train.py. The values of various hyperparameters are defined in `tools/options.py`.
### Inference
The trained models for partial experiments presented in the paper can be downloaded [here](https://drive.google.com/drive/folders/1inY29zupeCmvIF5SsJhQDEzo_jzY0j6Q). The inference script is `eval.py`. Here are some examples, assuming the models are downloaded in `logs`.
- AtLoc on `loop` from `RobotCar`:
```
python eval.py --dataset RobotCar --scene loop --model AtLoc --gpus 0 --weights ./logs/RobotCar_loop_AtLoc_False/models/epoch_300.pth.tar
```
Calculates the network attention visualizations and saves them in a video
- For the AtLoc model trained on `loop` in `RobotCar`:
```
python saliency_map.py --dataset RobotCar --scene loop --model AtLoc --gpus 0 --weights ./logs/RobotCar_loop_AtLoc_False/models/epoch_300.pth.tar
```
三维重建-利用注意力机制引导相机进行定位-优质项目实战.zip （102个子文件） 
ins.csv 20.21MB ins.csv 19.63MB ins.csv 18.93MB ins.csv 18.86MB vo.csv 3.88MB vo.csv 3.84MB vo.csv 3.81MB vo.csv 3.75MB ins.csv 1.86MB ins.csv 1.79MB ins.csv 1.64MB ins.csv 1.64MB gps.csv 1.45MB gps.csv 1.41MB gps.csv 1.36MB gps.csv 1.36MB vo.csv 383KB vo.csv 374KB vo.csv 355KB vo.csv 339KB gps.csv 137KB gps.csv 132KB gps.csv 121KB gps.csv 121KB .DS_Store 6KB
real.gif 13MB
trajectory.jpg 46KB README.md 3KB
mapnet-full1.png 71KB AtLocS-full1.png 66KB posenet-full1.png 65KB AtLoc-full1.png 63KB posenet-loop1.png 59KB posenet-loop2.png 59KB AtLoc-loop1.png 49KB AtLocS-loop1.png 48KB mapnet-loop1.png 47KB mapnet-loop2.png 46KB AtLoc-loop2.png 45KB Stereo-full1.png 45KB AtLocS-loop2.png 42KB Stereo-loop1.png 26KB Stereo-loop2.png 26KB dataloaders.py 12KB train.py 7KB utils.py 6KB eval.py 5KB saliency_map.py 4KB options.py 4KB atloc.py 4KB process_robotcar.py 2KB dataset_mean.py 2KB att.py 996B __init__.py 0B __init__.py 0B __init__.py 0B robotcar_symlinks.sh 396B stereo.timestamps 656KB stereo.timestamps 648KB stereo.timestamps 646KB stereo.timestamps 633KB stereo.timestamps 64KB stereo.timestamps 62KB stereo.timestamps 59KB stereo.timestamps 56KB
stats.txt 150B stats.txt 150B stats.txt 150B stats.txt 150B stats.txt 150B stats.txt 150B stats.txt 150B test_split.txt 86B pose_stats.txt 77B pose_stats.txt 76B train_split.txt 72B train_split.txt 61B stats.txt 60B stats.txt 60B pose_stats.txt 60B pose_stats.txt 60B pose_stats.txt 60B pose_stats.txt 60B test_split.txt 60B pose_stats.txt 60B pose_stats.txt 60B pose_stats.txt 60B train_split.txt 40B train_split.txt 40B test_split.txt 40B train_split.txt 40B train_split.txt 40B train_split.txt 40B test_split.txt 39B test_split.txt 20B test_split.txt 20B test_split.txt 20B train_split.txt 20B test_split.txt 20B train_split.txt 10B共 102 条
- 1
- 2
资源评论
