VIT track(gsoc realtime object tracking model) #24201
Vit tracker(vision transformer tracker) is a much better model for real-time object tracking. Vit tracker can achieve speeds exceeding nanotrack by 20% in single-threaded mode with ARM chip, and the advantage becomes even more pronounced in multi-threaded mode. In addition, on the dataset, vit tracker demonstrates better performance compared to nanotrack. Moreover, vit trackerprovides confidence values during the tracking process, which can be used to determine if the tracking is currently lost.
opencv_zoo: https://github.com/opencv/opencv_zoo/pull/194
opencv_extra: [https://github.com/opencv/opencv_extra/pull/1088](https://github.com/opencv/opencv_extra/pull/1088)
# Performance comparison is as follows:
NOTE: The speed below is tested by **onnxruntime** because opencv has poor support for the transformer architecture for now.
ONNX speed test on ARM platform(apple M2)(ms):
| thread nums | 1| 2| 3| 4|
|--------|--------|--------|--------|--------|
| nanotrack| 5.25| 4.86| 4.72| 4.49|
| vit tracker| 4.18| 2.41| 1.97| **1.46 (3X)**|
ONNX speed test on x86 platform(intel i3 10105)(ms):
| thread nums | 1| 2| 3| 4|
|--------|--------|--------|--------|--------|
| nanotrack|3.20|2.75|2.46|2.55|
| vit tracker|3.84|2.37|2.10|2.01|
opencv speed test on x86 platform(intel i3 10105)(ms):
| thread nums | 1| 2| 3| 4|
|--------|--------|--------|--------|--------|
| vit tracker|31.3|31.4|31.4|31.4|
preformance test on lasot dataset(AUC is the most important data. Higher AUC means better tracker):
|LASOT | AUC| P| Pnorm|
|--------|--------|--------|--------|
| nanotrack| 46.8| 45.0| 43.3|
| vit tracker| 48.6| 44.8| 54.7|
[https://youtu.be/MJiPnu1ZQRI](https://youtu.be/MJiPnu1ZQRI)
In target tracking tasks, the score is an important indicator that can indicate whether the current target is lost. In the video, vit tracker can track the target and display the current score in the upper left corner of the video. When the target is lost, the score drops significantly. While nanotrack will only return 0.9 score in any situation, so that we cannot determine whether the target is lost.
### Pull Request Readiness Checklist
See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request
- [x] I agree to contribute to the project under Apache 2 License.
- [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
- [x] The PR is proposed to the proper branch
- [ ] There is a reference to the original bug report and related work
- [x] There is accuracy test, performance test and test data in opencv_extra repository, if applicable
Patch to opencv_extra has the same branch name.
- [ ] The feature is well documented and sample code can be built with the project CMake
Fix python sample code (tst_scene_render) #24116
Fix bug of python sample code (samples/python/tst_scene_render.py) when backGr or fgr is None (#24114)
1) pass shape tuple to np.zeros arguments instead of integers
2) change np.int to int
### Pull Request Readiness Checklist
See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request
- [o] I agree to contribute to the project under Apache 2 License.
- [o] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
- [o] The PR is proposed to the proper branch
- [o] There is a reference to the original bug report and related work
- [o] There is accuracy test, performance test and test data in opencv_extra repository, if applicable
Patch to opencv_extra has the same branch name.
- [o] The feature is well documented and sample code can be built with the project CMake
- Fixed width and height swap in board size
- Fixed defaults in command line hint
- Fixed board visualization for Charuco case
- Used matchImagePoints method to handle partially detected Charuco boards
Added charuco pattern into calibrate.py #23587
### Pull Request Readiness Checklist
See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request
- [x] I agree to contribute to the project under Apache 2 License.
- [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
- [x] The PR is proposed to the proper branch
- [x] There is accuracy test, performance test and test data in opencv_extra repository, if applicable
Patch to opencv_extra has the same branch name.
- [x] The feature is well documented and sample code can be built with the project CMake
Add python sample of how to use Orbbec camera. #23531
### Pull Request Readiness Checklist
See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request
- [x] I agree to contribute to the project under Apache 2 License.
- [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
- [x] The PR is proposed to the proper branch
- [ ] There is a reference to the original bug report and related work
- [ ] There is accuracy test, performance test and test data in opencv_extra repository, if applicable
Patch to opencv_extra has the same branch name.
- [ ] The feature is well documented and sample code can be built with the project CMake
Usage of imread(): magic number 0, unchecked result
* docs: rewrite 0/1 to IMREAD_GRAYSCALE/IMREAD_COLOR in imread()
* samples, apps: rewrite 0/1 to IMREAD_GRAYSCALE/IMREAD_COLOR in imread()
* tests: rewrite 0/1 to IMREAD_GRAYSCALE/IMREAD_COLOR in imread()
* doc/py_tutorials: check imread() result
[teset data in opencv_extra](https://github.com/opencv/opencv_extra/pull/1016)
NanoTrack is an extremely lightweight and fast object-tracking model.
The total size is **1.1 MB**.
And the FPS on M1 chip is **150**, on Raspberry Pi 4 is about **30**. (Float32 CPU only)
With this model, many users can run object tracking on the edge device.
The author of NanoTrack is @HonglinChu.
The original repo is https://github.com/HonglinChu/NanoTrack.
### Pull Request Readiness Checklist
See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request
- [x] I agree to contribute to the project under Apache 2 License.
- [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
- [x] The PR is proposed to the proper branch
- [ ] There is a reference to the original bug report and related work
- [ ] There is accuracy test, performance test and test data in opencv_extra repository, if applicable
Patch to opencv_extra has the same branch name.
- [ ] The feature is well documented and sample code can be built with the project CMake
### Critical bugs fixed:
- `seam_finder.find()` returns None and overwrites `masks_warped`
- `indices` is only 1-dimensional
### Nice-to-have bugs fixed:
- avoid invalid value in sqrt and subsequent runtime warning
- avoid printing help string on each run (use argparse builtin behavior)
### New features:
- added graphcut seam finder support
### Test Summary:
Tested on Ubuntu 20.04 with python 3.8.10 and opencv-python-contrib 4.5.5.62
fix cvtColor-error
* fix gray image channel error
* fix gray image channel error
* fix cvtColor error after the video end
* fix cvtColor error after the video end and change next variable
* fix cvtColor error after the video end
* reset next variable
* fix cvtColor error after the video end
* fix cvtColor error after the video end
without rounding the composed image sizes (variable "sz") they will be odly fractions of a pixel (e.g. (5300.965, 3772.897)) and therefore cause a "TypeError: integer argument expected, got float" in line
456 roi = warper.warpRoi(sz, K, cameras[i].R)
Stitching Detailed Tutorial Improvements
* Add Vertical Wave Correction
The user has the possibility to pass "vert" as wave_correct parameter. However, in the code "cv.detail.WAVE_CORRECT_HORIZ" ist fixed. This change proposes changes so that the wave correction is done vertically if the user passes "vert" as wave_correct parameter. The variable "do_wave_correct" is replaced by None which is passed to the variable "wave_correct" if the user chooses "no" for wave correction.
* Correct fixed conf_thresh
According to the documentation, [cv.detail.leaveBiggestComponent](https://docs.opencv.org/4.5.1/d7/d74/group__stitching__rotation.html#ga855d2fccbcfc3b3477b34d415be5e786) takes features, the pairwise_matches and the conf_threshold as input.
In the tutorial, however, conf_threshold is fixed at 0.3 even though the user can pass conf_thresh as parameter which is 1 by default. Fixing this parameter at 0.3 causes the script to include images into the panorama which are not part of it.
* Error Message for SURF if not implemented
In OpenCV 4.5.1
import cv2 as cv
cv.xfeatures2d_SURF.create
will not create an AttributeError, even if the function is excluded (no nonfree option)
In Line 305 (now 306) however ´finder = FEATURES_FIND_CHOICES[args.features]()´ will raise an
error: OpenCV(4.5.1) ..\opencv_contrib\modules\xfeatures2d\src\surf.cpp:1029: error: (-213:The function/feature is not implemented) This algorithm is patented and is excluded in this configuration; Set OPENCV_ENABLE_NONFREE CMake option and rebuild the library in function 'cv::xfeatures2d::SURF::create'
So we should check with cv.xfeatures2d_SURF.create() correctly if SURF is available
Instead of being a copy of line 76, line 79 instead correctly indicates that it will show a histogram for a gray image in curve mode, as given by the code block at line 103 referencing image "gray" instead of image "im".
