shiaoming / python-vo Goto Github PK
View Code? Open in Web Editor NEWA simple python implemented frame-by-frame visual odometry with SuperPoint feature detector and SuperGlue feature matcher.
A simple python implemented frame-by-frame visual odometry with SuperPoint feature detector and SuperGlue feature matcher.
您好,运行程序后,有以下两个问题想请教您:
1. 最后生成的AvgError的图只有一个红点,并未有完整的轨迹,是什么问题?
2. 我是否可以用我自己的图片在此程序上进行图像匹配,请问groundtruth文件怎样能获得?
您好,请问有尝试复现superglue文章里面的sift+superglue吗?最近一直在尝试,一直没有进展,希望可以交流
Hello, may I ask, how to output the matching results of image feature points?
I try to run code in TUM datasets but get the error message.
Traceback (most recent call last):
File "main.py", line 109, in
run(args)
File "main.py", line 65, in run
loader = create_dataloader(config["dataset"])
File "/home/zhy/PyProject/Python-VO/DataLoader/init.py", line 9, in create_dataloader
loader = eval(code_line)
File "", line 1, in
File "/home/zhy/PyProject/Python-VO/DataLoader/TUMRGBLoader.py", line 32, in init
self.read_imgs()
File "/home/zhy/PyProject/Python-VO/DataLoader/TUMRGBLoader.py", line 76, in read_imgs
rgb_list = lines[:, 1]
IndexError: too many indices for array
I use this code for running:
python main.py --config params/tumrgb_superpoint_supergluematch.yaml
The TUM dataset is organized as follows:
├ TUM
└rgbd_dataset_freiburg1_360
└── depth
└ ...
└── rgb
└ ...
└── depth.txt
└── rgb.txt
└── rgbd_gt.txt
How to fix this error?
orb特征点的结果真的有这么差吗,01序列甚至无法完全跑完就报错了
When I try running the result_eval.py, I get the following error:
result_eval.py:21: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.
return np.array(img_ids), np.array(est_points), np.array(gt_points)
Traceback (most recent call last):
File "result_eval.py", line 70, in
main()
File "result_eval.py", line 38, in main
errors = np.linalg.norm((gt_xyz - est_xyz), axis=1)
ValueError: operands could not be broadcast together with shapes (8480,) (8480,3)
when i run main.py in TUM dataset, only one red dot is shown in the plot, and the trajectory is not shown.
thank you!
Thank you for your good work. I encountered this problem when used
$ python main.py
Traceback (most recent call last):
File "main.py", line 124, in
run(args)
File "main.py", line 110, in run
cv2.imwrite("results/" + fname + '_new.png', img2)
UnboundLocalError: local variable 'img2' referenced before assignment
then,I put cv2.imwrite() into for()
it works ,but the four methods all just load the model,like this
INFO:root:{'descriptor_dim': 256, 'weights': 'outdoor', 'keypoint_encoder': [32, 64, 128, 256], 'GNN_layers': ['self', 'cross', 'self', 'cross', 'self', 'cross', 'self', 'cross', 'self', 'cross', 'self', 'cross', 'self', 'cross', 'self', 'cross', 'self', 'cross'], 'sinkhorn_iterations': 100, 'match_threshold': 0.2, 'cuda': 1, 'type': 'SuperGlue'}
INFO:root:creating SuperGlue matcher...
Loaded SuperGlue model ("outdoor" weights)
Thank you.
can this part be more specific?
您好,我在运行的时候,会报错UnboundLocalError: local variable 'img2' referenced before assignment,我在前面加了global img2 也没用,我的Python版本是3.7,请问您知道怎么解决吗?
我想问一下这个AvgError的精度跟显存有关系吗?我的是TITAN XP 12G的显存
hi,progessor:
like the topic said,please help!
How do i convert the matched keypoints to 3d landmarks after apploying superglue matching ? Such that they can be converted into pointcloud and saved to database?
RuntimeError: CUDA out of memory. Tried to allocate 20.00 MiB (GPU 0; 6.00 GiB total capacity; 4.08 GiB already allocated; 10.56 MiB free; 4.18 GiB reserved in total by PyTorch)
请问有试过在TUM RGBD fr3 long office那个序列上跑吗?
我跑出来结果非常差,各种方法都得不到一个合理的轨迹,不知道是不是存在问题还是这个序列本身太难。
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.