从相机显示具有特定旋转的完整对象C＃

Question

I have different kind of objects with variable dimensions and placed at different position in a scene. I want to focus/display each object with the camera with a hard code rotation (north rotation). Like with specific camera rotation I want to focus the object that it completely show to the camera and center of the screen. For this reason, I have write this code snippet that 1

1. Get the position behind the specific focus object using TransformPoint

2. Provide the elevation using the largest extent of the bound so that it display maximum area of the object.

3. Assign the position and the fixed rotation

     Vector3 dest = destination.transform.TransformPoint(0, 0, behindPositionDistance);// GetBehindPosition(destination.transform, behindPositionDistance, elevation); Debug.Log(dest); float eleveMax = Mathf.Max(destination.GetComponent<MeshRenderer>().bounds.extents.x, destination.GetComponent<MeshRenderer>().bounds.extents.z); dest = new Vector3(dest.x, eleveMax, dest.z); camera.transform.position = dest; camera.transform.rotation = lookNorth; 

But the problem is, it is not accurately working with all objects as every object is at different position and dimension. I want to focus the full object using the camera but without changing the rotation.

Answer 1

Maybe create an empty gameobject as a child of your objects. Use that as your Camera position at runtime.

Answer 2

If you can use an orthographic camera, you can set the camera's orthographicSize dynamically based on the size of the bounds of the GameObject. This is probably the easiest solution.

If you need a perspective camera, you can get the Plane s or corners of the camera frustum via GeometryUtility.CalculateFrustumPlanes(Camera.main) or Camera.CalculateFrustumCorners and use the results from either to manually test that your Bounds is entirely inside.

With a perspective camera, you could also compute the necessary distance from the object based on the size of the object's bounds and the FOV of the camera. If I'm not mistaken, this would be something along the lines of distance = 0.5 * size / tan(0.5 * fov) , but the function doesn't have to be precise; it just needs to work for your camera. Alternatively, you could keep distance constant and compute FOV from the size of the object, although I wouldn't recommend that because frequent FOV changes sound disorienting for the viewer; my point is that there are many options.