I am currently facing a unique challenge:
In my project, I have markers (specifically Hiro, A, Kanji) displayed on banners, and I need to show a 3D model (such as augmented reality) when the marker is recognized. Everything works well so far, but here's the catch - the 3D model needs to be anchored to the ground. The ground is not perfectly flat or extremely bumpy; it resembles a meadow.
My question is: Is it possible to anchor the 3D model to the ground without using any physical tools (I mean not involving hardware, only software)?
If yes, what do you think would be the most straightforward method? I've heard that using OpenCV.js might provide a solution, but I haven't come across any useful resources yet.
The 3D model appears correctly on the marker; all I need to do is determine its positioning so that it sits precisely on the ground. But how can I find these x, y, z coordinates?
Currently, I don't have any code to share because I'm unsure where to begin in order to make my work effective.
The entire process, starting from scanning the marker through the webcam, takes place on a website. I am utilizing NodeJS & ExpressJS on the backend, EJS as the template language, and THREE.js for the augmented reality aspect.
So, do you believe there is a way to accomplish this task?