AR App：拍摄图案并将其转换为图像平面
我正在做增强现实研究并使用 OpenCV。我已经找到了棋盘格，并且想在棋盘格上放置一些 3D 图形。
在 OpenCV 中，findCheckerboard 函数将找到棋盘格块的角点。到目前为止，我所做的就是获取四个最外角并将图像放置在棋盘上进行仿射变换。
我想将其扩展到 3D 模型。我认为，一旦获得四个角，您就会以某种方式移动相机，使其看起来像 3D 图像位于棋盘顶部。我对这个问题的思考正确吗？
如果是这样，你到底是怎么做到的？
I'm doing Augmented Reality research and am using OpenCV. I have the find checkerboard working and am wanting to put some 3D graphics on the checkerboard.
In OpenCV the findCheckerboard function will find the corners of the blocks of your checkerboard. What I have done so far is taken the four outermost corners and placed an image on the checkerboard doing an affine transformation.
I'm wanting to expand this to a 3D model. I'm thinking that once you get the four corners you'll somehow move the camera to make it look like the 3D image is on top of the checkboard. I'm I thinking correctly about this?
If so, how on earth do you do that?
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
我建议使用 osgART 因为它包含 OpenSceneGraph 这将有助于将 3D 模型附加到您的标签。除非您确实需要使用 OpenCV，否则使用 OpenCV 进行所有检测、定向和定位将比仅使用 AR 库多做一些工作。
如果 osgART 不适合您，请查看 ARToolKit 或其衍生物(ARToolKitPlus) 以获得更适合您需求的东西。
I'd suggest using osgART for this since it contains OpenSceneGraph which would help with attaching 3D models to your tag(s). Unless you really need to use OpenCV, doing all the detection, orientation, and positioning with OpenCV would be a bit more work than just using an AR library.
If osgART isn't for you, look into ARToolKit or its derivatives(ARToolKitPlus) for something that better suits your needs.