How to know if an object is on top of another object?

I'm creating a 2D puzzle. Here, the rule is that, player clicks on puzzle pieces. If the piece is not under another piece, it will be cleared. However, if there is another piece on top of it, a penalty is applied.

One approach i thought was to use renderer.SortingOrder. But, problem is, it is common for two a higher order piece to exist in the scene and not be on top of a lower order object. In such a case, the lower order piece can be removed.
For example, this image:

Here, yellow is on top of blue, blue is on top of green and green is on top of red. So to solve the puzzle, the player has to first remove yello, then blue,green and finally red. Now if there was another orange piece not above or below any of these pieces, it can be removed first.

 

Raycasting is another idea that i considered. But it works only if center of the object is overlapped by another object.

 

You could always try adding triggers to each side of the blocks.

 

Using colliders that are "higher" than your planes, and checking if they are colliding with another of these colliders, and if so, check if the other object is below or above your plane. I think that's the best idea

 

You could simply keep track of the positions using your own depth order of sorts. If you are using a 3d view(even if 2d game) you can simply use an axis position, say if they are further in the z direction. And if you are using an ortho camera things won't get smaller no matter how far away they are.

I assume you are already in some fashion sorting the objects by distance(or something) because they are getting rendered that way. Just use the same system.

 

@N1warhead , it's 2D with Orthographic camera. I don't think i can add trigger to both sides of the blocks.

@kburkhart84 , with that problem is, it is possible say one block has depth order 1 and another 0. Now, it is possible that both these objects are not in contact with each other. In that case, either of them can be cleared any time.

 

If there is only ever 1 top-most item, you can check for clicks on just that object. If you receive a click event on an object that is not that one, apply your penalty. When you do click on the top one, remove it and set the next most-top one to be checked.

 

That shouldn't matter. Unity is always 3d, the Camera is just Orthographic, that's it. It just puts your sprites across world space coordinates;

I've never done 2D, but what I said above is true;

 

@Tomnnn , there can be multiple topmost items.

@N1warhead , thanks i'll implement that method and see.

Hmmm i can't use rigidbody. So can't use triggers

 

Is the issue choosing & determining the top or having them visually in order? If you're using an orthographic camera and you use 3D planes for the objects, you can just order their transforms. Since it's an orthographic camera, the sizes will never change, but the visual layering should. Then a raycast should work, and you can check the z distance from the camera to determine if it is a top item.

 

@Tomnnn the problem is, it is possible that two pieces of varying transforms are on top. For example, in the image i have attached, say there is an orange block with same transforms ordering as the blue block. However, this orange block is placed separately without being in contact with any other piece. In such a case, blue block cannot be removed unless yellow is removed. But, orange block can be removed at any time since it is not in contact with any other block.

 

Isn't that intended? I was already assuming your game was being inspired by mahjong

What you're describing sounds like an issue with design. If it isn't intended for the orange block to be removed yet, why is it placed in the same layer as the blue block with nothing above it? What is supposed to happen? If your game is not supposed to have cases like this, do not consider them while designing your other systems. Focus on making sure your level generator does not bug out and produce incorrect layouts

If this kind of situation is intended to happen, how is the user supposed to handle it?

 

Sorry, i didnt really understand your question.

 

@Tomnnn this is embarrassing :x i did not know how Mahjong is played. Since you brought up the similarity, it's easier to find the required solution. Thanks a lot.

 

Oh, haha. Is it similiar? Good luck then ^-^

 

I would just set the Z order of all objects based on the order they are layered, then with a raycast again triggers/colliders, if the player chooses one with a Z order that is `further away` than the closest remaining Z coordinate, it's an error. Keep track of the closest object's Z coord, use an array of object positions. etc

The game is basically `pick up sticks`

 

You need to use a sweep rather than a ray, though from memory that only works for certain shapes.

Edit: Checked the docs. Works for anything that's a Rigidbody as long as it's not using a mesh collider. You might also have to do an overlap test if there's any possibility that your objects can penetrate one another, as I don't know if the sweep will report pre-existing collisions.

If there are non-overlapping stacks of different heights that won't work, though.

 

Why not use Physics2D.OverlapAreaAll to find all the overlapping sticks?

Here is a working example:

Code (CSharp):

1. [RequireComponent(typeof(BoxCollider2D))]
2. public class Stick : MonoBehaviour
3. {
4.     void OnMouseDown()
5.     {
6.         // Find all colliders that overlap
7.         BoxCollider2D myCollider = GetComponent<BoxCollider2D>();
8.         Collider2D[] otherColliders = Physics2D.OverlapAreaAll(myCollider.bounds.min, myCollider.bounds.max);
9.  
10.         // Check for any colliders that are on top
11.         bool isUnderneath = false;
12.         foreach (var otherCollider in otherColliders)
13.         {
14.             if (otherCollider.transform.position.z < this.transform.position.z)
15.             {
16.                 isUnderneath = true;
17.                 break;
18.             }
19.         }
20.  
21.         // Take the appropriate action
22.         if (!isUnderneath)
23.         {
24.             Debug.Log("HOORAY!");
25.             Destroy(this.gameObject);
26.         }
27.         else
28.         {
29.             Debug.Log("OOPS!");
30.         }
31.     }
32. }

 

@AdamScura Thanks a lot. I'll try it.
@angrypenguin i can't use rigidbody in this case.
@imaginaryhuman i checked pickup sticks. Except for the pairing part, it does look similar. Thanks.

 

Why not?

 

2D game, orthographic camera. Don't need rigidbody. If i make them rigidbody, they will all fall down

 

That just means the Rigidbody isn't set up properly.

Also, I haven't looked into the 2D stuff much so it could be different, but it doesn't matter what type of game it is or what camera you have - if you've got an object with collider and it moves it should have a Rigidbody. Check the docs. Don't be fooled by the fact that the default Rigidbody config uses physics and gravity.

 

My objects don't move

 

Do they spawn/unspawn at runtime? Same thing.

Regardless, an existing solution has been demonstrated using only a built-in component. It's no pain to anyone else if you choose not to use it.

 

Oh sorry i didn't pay attention to the sweep part. That certainly looks like the easiest way.
All i need to do is check if it fits in it's current.Z+1. Thanks

 

Haha, no problem and you're welcome.

 

Now that I think about it, @angrypenguin's approach is simpler than mine.

Here is the code:

Code (CSharp):

1.  
2. [RequireComponent(typeof(Rigidbody))]
3. [RequireComponent(typeof(BoxCollider))]
4. public class Stick3D : MonoBehaviour
5. {
6.     void OnMouseDown()
7.     {
8.         RaycastHit hit;
9.         rigidbody.SweepTest(Vector3.back, out hit);
10.         if (hit.collider == null)
11.         {
12.             Debug.Log("HOORAY!");
13.             Destroy(this.gameObject);
14.         }
15.         else
16.         {
17.             Debug.Log("OOPS!");
18.         }
19.     }
20. }
21.  

Enable "Is Kinematic" on the Rigidbody if you don't want them to move.

 

You don't even have to check the hit collider. It also returns a bool, so you could do this...

Code (csharp):

1. bool didHit = rigidbody.SweepTest(Vector3.back, out hit);
2. if (!didHit)
3. {
4.     Debug.Log("HOORAY!");
5.     Destroy(this.gameObject);
6. }
7. else
8. {
9.     Debug.Log("OOPS!");
10. }

What was your other solution? (I remember looking at it and not thinking there was anything wrong with it.)

Also disable gravity.

 

Nice.

My first solution was to use BoxCollider2D and Physics2D.OverlapAreaAll. Yours is more simple because it does not require a for loop. IMO SweepTest is the way to go.

 

Thank you so much. I was stuck going nowhere on this

 

Hope it helps!

 

Tried this script and it didn't work as expected.

void Update () {

if (Input.GetMouseButtonDown(0)) {
Debug.Log("Pressed left click, casting ray.");
CastRay();
}
}
void CastRay() {
Ray ray = Camera.main.ScreenPointToRay (Input.mousePosition);
RaycastHit hit,hit1;
if (Physics.Raycast (ray, out hit, 100)) {
didHit=GameObject.Find (hit.collider.name).GetComponent<Rigidbody>().SweepTest(Vector3.back,out hit1);

if(didHit)Debug.Log ("Nope. Can't");
else
Destroy(GameObject.Find (hit.collider.name));
}
}

For this, arranged objects as shown in the image:


The horizontal bar is under the vertical bars and should not be destroyed. But, it is getting destroyed.

However, i can combine ideas provided here to get a work around. One way would be to attach triggers, then transform.position.back to see if trigger is fired.

Thanks for helping

 

Hmmm it worked like a charm when i used -Vector3.forward instead of vector3.back. Thank you. That was a big help.

 

That's really bizarre. -Vector3.forward == Vector3.back. Make sure your pieces are thin enough that they don't intersect each other on the Z-Axis or you will get unpredictable results. (example: use a z-spacing of 1 and a BoxCollider thickness of 0.9)

FYI: You can simplify your code by taking advantage of OnMouseDown(). When the user clicks the mouse, Unity does a camera-raycast internally and only calls OnMouseDown() on the closest clicked GameObject. That means you don't need to do your own raycast if you use OnMouseDown(). See my previous example.

Happy Coding!

 

With that approach, i'd have to attach a script to every piece. That's why i'm using my own raycasting