file-type

相机标定的新方法:灵活技术详解

PDF文件

下载需积分: 14 | 2.33MB | 更新于2024-07-16 | 150 浏览量 | 0 下载量 举报 收藏
download 立即下载
“张正友的相机标定技术” 这篇由微软研究院的 Zhengyou Zhang 编写的报告“一种灵活的新相机标定技术”是关于计算机视觉领域的重要贡献,它提出了一个创新的方法来解决相机校准问题。相机标定是计算摄影学中的核心任务,对于精确的三维重建和图像处理至关重要。 报告首先介绍了研究的动机。在计算机视觉系统中,由于镜头的光学特性、传感器的非线性以及相机与被摄物体间的相对位置变化,图像会存在失真。这种失真影响了对真实世界场景的准确理解,因此需要通过标定来校正。 接着,Zhang 描述了基本的数学方程。他定义了符号和术语,并讨论了模型平面与其图像之间的同构关系。同构关系允许将三维空间的几何信息映射到二维图像平面上。他还探讨了内在参数的约束,这些参数包括焦距、主点坐标和光轴方向等。 在解决相机标定问题时,Zhang 提出了两种方法:闭式解和最大似然估计。闭式解提供了一种直接求解相机参数的公式化方法,而最大似然估计则通过优化数据拟合度来提高标定的准确性。这两种方法都是为了在考虑实际的随机噪声和系统误差情况下,找到最符合观测数据的相机参数。 报告中还涉及了如何处理径向畸变,这是大多数镜头常见的失真类型,使得远离或靠近图像中心的物体变形。Zhang 的方法能够校正这种失真,从而提高图像恢复的准确性。 进一步,Zhang 讨论了可能的退化配置,即某些特定的场景设置可能导致标定问题变得难以解决或无解。理解这些退化情况有助于设计更稳健的标定算法。 实验结果部分展示了该方法在计算机模拟和真实数据上的性能。通过比较不同条件下的标定结果,包括模型点的随机噪声和模型模式的非平面性,Zhang 验证了该方法的稳定性和鲁棒性。 这篇技术报告提供了一种灵活且实用的相机标定技术,不仅解决了传统的相机标定难题,还为后续的计算机视觉研究提供了坚实的理论基础。这一工作对于自动驾驶、无人机导航、虚拟现实等依赖精确图像理解的应用具有深远的影响。

相关推荐

filetype

# 3D Cinemagraphy from a Single Image (CVPR 2023) [Xingyi Li](https://scholar.google.com/citations?user=XDKQsvUAAAAJ&hl)1,3, [Zhiguo Cao](http://english.aia.hust.edu.cn/info/1085/1528.htm)1, [Huiqiang Sun](https://huiqiang-sun.github.io/)1, [Jianming Zhang](https://jimmie33.github.io/)2, [Ke Xian](https://kexianhust.github.io/)3*, [Guosheng Lin](https://guosheng.github.io/)3 1Huazhong University of Science and Technology, 2Adobe Research, 3S-Lab, Nanyang Technological University [Project](https://xingyi-li.github.io/3d-cinemagraphy/) | [Paper](https://github.com/xingyi-li/3d-cinemagraphy/blob/main/pdf/3d-cinemagraphy-paper.pdf) | [arXiv](https://arxiv.org/abs/2303.05724) | [Video](https://youtu.be/sqCy7ffTEEY) | [Supp](https://github.com/xingyi-li/3d-cinemagraphy/blob/main/pdf/3d-cinemagraphy-supp.pdf) | [Poster](https://github.com/xingyi-li/3d-cinemagraphy/blob/main/pdf/3d-cinemagraphy-poster.pdf) This repository contains the official PyTorch implementation of our CVPR 2023 paper "3D Cinemagraphy from a Single Image". ## Installation ``` git clone https://github.com/xingyi-li/3d-cinemagraphy.git cd 3d-cinemagraphy bash requirements.sh ``` ## Usage Download pretrained models from [Google Drive](https://drive.google.com/file/d/1ROxvB7D-vNYl4eYmIzZ5Gitg84amMd19/view?usp=sharing), then unzip and put them in the directory `ckpts`. To achieve better motion estimation results and controllable animation, here we provide the controllable version. Firstly, use [labelme](https://github.com/wkentaro/labelme) to specify the target regions (masks) and desired movement directions (hints): ```shell conda activate 3d-cinemagraphy cd demo/0/ labelme image.png ``` A screenshot here: ![labelme](assets/labelme.png) It is recommended to specify **short** hints rather than long hints to avoid artifacts. Please follow [labelme](https://github.com/wkentaro/labelme) for detailed instructions if needed. After that, we can obtain an image.json file. Our next step is to convert the annotations stored in JSON format into datasets that can be used by our method: ```shell labelme_json_to_dataset image.json # this will generate a folder image_json cd ../../ python scripts/generate_mask.py --inputdir demo/0/image_json ``` We now can create 3D cinemagraphs according to your preference: ```shell python demo.py -c configs/config.yaml --input_dir demo/0/ --ckpt_path ckpts/model_150000.pth --flow_scale 1.0 --ds_factor 1.0 ``` - `input_dir`: input folder that contains src images. - `ckpt_path`: checkpoint path. - `flow_scale`: scale that used to control the speed of fluid, > 1.0 will slow down the fluid. - `ds_factor`: downsample factor for the input images. Results will be saved to the `input_dir/output`. ## Known issues - Due to the limited size of the training dataset, the intermediate frame may occasionally experience flickering. - The utilization of a fixed distance threshold in agglomerative clustering within the disparity space can occasionally result in the presence of visible boundaries between different layers. - We may sometimes see artifacts when the fluid is moving very fast. You can either slow down the fluid by increasing the `flow_scale` or try to specify short hints rather than long hints, to avoid artifacts. - The motion estimation module occasionally generates motion fields that do not perfectly align with the desired preferences. ## Citation If you find our work useful in your research, please consider to cite our paper: ``` @InProceedings{li2023_3dcinemagraphy, author = {Li, Xingyi and Cao, Zhiguo and Sun, Huiqiang and Zhang, Jianming and Xian, Ke and Lin, Guosheng}, title = {3D Cinemagraphy From a Single Image}, booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)}, month = {June}, year = {2023}, pages = {4595-4605} } ``` ## Relevant works - [Animating Pictures with Eulerian Motion Fields](https://openaccess.thecvf.com/content/CVPR2021/papers/Holynski_Animating_Pictures_With_Eulerian_Motion_Fields_CVPR_2021_paper.pdf), CVPR 2021 - [Controllable Animation of Fluid Elements in Still Images](https://openaccess.thecvf.com/content/CVPR2022/papers/Mahapatra_Controllable_Animation_of_Fluid_Elements_in_Still_Images_CVPR_2022_paper.pdf), CVPR 2022 - [Simulating Fluids in Real-World Still Images](https://arxiv.org/pdf/2204.11335), arXiv 2022 - [3D Photography using Context-aware Layered Depth Inpainting](https://openaccess.thecvf.com/content_CVPR_2020/papers/Shih_3D_Photography_Using_Context-Aware_Layered_Depth_Inpainting_CVPR_2020_paper.pdf), CVPR 2020 - [3D Photo Stylization: Learning to Generate Stylized Novel Views from a Single Image](https://openaccess.thecvf.com/content/CVPR2022/papers/Mu_3D_Photo_Stylization_Learning_To_Generate_Stylized_Novel_Views_From_CVPR_2022_paper.pdf), CVPR 2022 - [3D Moments from Near-Duplicate Photos](https://openaccess.thecvf.com/content/CVPR2022/papers/Wang_3D_Moments_From_Near-Duplicate_Photos_CVPR_2022_paper.pdf), CVPR 2022 - [3D Video Loops from Asynchronous Input](https://openaccess.thecvf.com/content/CVPR2023/papers/Ma_3D_Video_Loops_From_Asynchronous_Input_CVPR_2023_paper.pdf), CVPR 2023 ## Acknowledgement This code borrows heavily from [3D Moments](https://github.com/google-research/3d-moments) and [SLR-SFS](https://github.com/simon3dv/SLR-SFS). We thank the respective authors for open sourcing their methods.

江南又旧雨
  • 粉丝: 196
上传资源 快速赚钱