How to make fill method for photo editor? - java

I am making a pixel art editor in java, just for fun, and I ran into a problem. the problem occurred when I was trying to make the fill function. here is the code
private void fill(int x, int y){
Color beforeColor = img[x][y];
img[x][y] = foregroundColor;
if(x-1 >= 0){
if(img[x-1][y] == beforeColor){
fill(x-1, y);
}
}
if(x+1 >= 0){
if(img[x+1][y] == beforeColor){
fill(x+1, y);
}
}
if(y-1 >= 0){
if(img[x][y-1] == beforeColor){
fill(x, y-1);
}
}
if(y+1 >= 0){
if(img[x][y+1] == beforeColor){
fill(x, y+1);
}
}
}
img is an array of awt Color objects.
this method basically checks around the specified pixel for pixels of the same colour and then runs the method again for the next and the next until the whole area get filled.
if you know anything about the stack in the computer and recursion, then you'll probably realise that this will quickly cause a stackoverflowerror and halt the program. what I am trying to figure out is a way around recursion. Could someone please point me in the right direction around recursion and stackoverflowerror? thanks in advance to people who can help.

Might be that you could use a list: initially you fill it with the pixel that you want to start with. Then you have a loop that iterates until the list is empty. In the loop you take one pixel from the list and check its color. If the color matches recolor it and its neightbors to the list. If the color does not match ignore the pixel. So I would use a little helper class that stores the coordinates of a pixel. So it could look like that (some error checks,... might be missing, but that might be a way to go):
class PixelCoordinate {
public int x;
public int y;
public PixelCoordinate(int x, int y) {
this.x = x; this.y = y;
}
}
Color beforeColor = img[x][y];
List<PixelCoordinate> worklist = new ArrayList<PixelCoordinate>();
// The pixel to start with
worklist.add(new PixelCoordinate(x, y));
while (worklist.isEmpty() == false) {
// Take one pixel from the list
PixelCoordinate pixel = list.get(0);
list.remove(0);
// Check its color
if (img[x][y].equals(beforeColor) {
// Apply new color
img[x][y] = foregroundColor;
// Check neighbors
if (x-1 >= 0) {
list.add(new PixelCoordinate(x-y, y));
}
// Add other neighbors...
}
}

Related

How should I calculate every possible endpoint if my player can end on x movements in a turn-based game?

I'm currently recreating a Civilization game in Processing. I'm planning to implement the feature in a which a given unit can see every possible move it can make with a given number of hexes it is allowed to move. All possible endpoints are marked with red circles. However, units cannot move through mountains or bodies of water. I'm trying to approach this by finding out every possible combination of moves I can make without the unit going into a mountain or body of water but I can't figure out how I can determine every combination.
There are 6 directions that any unit can go in, north-east, north, north-west, south-east, south, south-west. The max number of movements I'm assigning to any unit would probably go up to 6. Any higher and I'm afraid processing may become to slow every time I move a unit.
I'm trying to recreate this:
What I'm hoping the result will look like with two possible movements (without the black arrows):
Raw version of that image:
Here is the code I use to draw the hex grid. Immediately after drawing each individual hex, its center's x coords and y coords are stored in xHexes and yHexes respectively. Also, immediately after generating the type of tile (e.g. grass, beach), the type of tile is also stored in an array named hexTypes. Therefore, I can get the x and y coords and type of hex of any hex I want on the map just by referencing its index.
Code used to draw a single hexagon:
beginShape();
for (float a = PI/6; a < TWO_PI; a += TWO_PI/6) {
float vx = x + cos(a) * gs*2;
float vy = y + sin(a) * gs*2;
vertex(vx, vy);
}
x is the x coord for centre of hexagon
y is the y coord for centre of hexagon
gs = radius of hexagon
Code used to tesselate hex over the window creating a hex grid:
void redrawMap() {
float xChange = 1.7;
float yChange = 6;
for (int y = 0; y < ySize/hexSize; y++) {
for (int x = 0; x < xSize/hexSize; x++) {
if (x % 2 == 1) {
// if any part of this hexagon being formed will be visible on the window and not off the window.
if (x*hexSize*xChange <= width+2*hexSize && int(y*hexSize*yChange) <= height+3*hexSize) {
drawHex(x*hexSize*xChange, y*hexSize*yChange, hexSize);
}
// only record and allow player to react with it if the entire tile is visible on the window
if (x*hexSize*xChange < width && int(y*hexSize*yChange) < height) {
xHexes.add(int(x*hexSize*xChange));
yHexes.add(int(y*hexSize*yChange));
}
} else {
if (x*hexSize*xChange <= width+2*hexSize && int(y*hexSize*yChange) <= height+3*hexSize) {
drawHex(x*hexSize*xChange, y*hexSize*yChange+(hexSize*3), hexSize);
}
if (x*hexSize*xChange < width && int(y*hexSize*yChange+(hexSize*3)) < height) {
xHexes.add(int(x*hexSize*xChange));
yHexes.add(int(y*hexSize*yChange+(hexSize*3)));
}
}
}
}
}
hexSize is a user-specified size for each hexagon, determining the number of hexagons that will be on the screen.
This answer will help you get to this (green is plains, red is hills and blue is water, also please don't flame my terrible grid):
Note that there is no pathfinding in this solution, only some very simple "can I get there" math. I'll include the full code of the sketch at the end so you can reproduce what I did and test it yourself. One last thing: this answer doesn't use any advanced design pattern, but it assume that you're confortable with the basics and Object Oriented Programming. If I did something which you're not sure you understand, you can (and should) ask about it.
Also: this is a proof of concept, not a "copy and paste me" solution. I don't have your code, so it cannot be that second thing anyway, but as your question can be solved in a bazillion manners, this is only one which I deliberately made as simple and visual as possible so you can get the idea and run with it.
First, I strongly suggest that you make your tiles into objects. First because they need to carry a lot of information (what's on each tile, how hard they are to cross, maybe things like resources or yield... I don't know, but there will be a lot of stuff).
The Basics
I organized my global variables like this:
// Debug
int unitTravelPoints = 30; // this is the number if "travel points" currently being tested, you can change it
// Golbals
float _tileSize = 60;
int _gridWidth = 10;
int _gridHeight = 20;
ArrayList<Tile> _tiles = new ArrayList<Tile>(); // all the tiles
ArrayList<Tile> _canTravel = new ArrayList<Tile>(); // tiles you can currently travel to
The basics being that I like to be able to change my grid size on the fly, but that's just a detail. What's next is to choose a coordinate system for the grid. I choose the simplest one as I didn't want to bust my brain on something complicated, but you may want to adapt this to another coordinate system. I choose the offset coordinate type of grid: my "every second row" is half a tile offset. So, instead of having this:
I have this:
The rest is just adjusting the spatial coordinates of the tiles so it doesn't look too bad, but their coordinates stays the same:
Notice how I consider that the spatial coordinates and the grid coordinates are two different things. I'll mostly use the spatial coordinates for the proximity checks, but that's because I'm lazy, because you could make a nice algorithm which do the same thing without the spatial coordinates and it would probably be less costly.
What about the travel points? Here's how I decided to work: your unit has a finite amount of "travel points". Here there's no unit, but instead a global variable unitTravelPoints which will do the same thing. I decided to work with this scale: one normal tile is worth 10 travel points. So:
Plains: 10 points
Hills: 15 points
Water: 1000 points (this is impassable terrain but without going into the details)
I'm not going to go into the details of drawing a grid, but that's mostly because your algorithm looks way better than mine on this front. On the other hand, I'll spend some time on explaining how I designed the Tiles.
The Tiles
We're entering OOP: they are Drawable. Drawable is a base class which contains some basic info which every visible thing should have: a position, and an isVisible setting which can be turned off. And a method to draw it, which I call Render() since draw() is already taken by Processing:
class Drawable {
PVector position;
boolean isVisible;
public Drawable() {
position = new PVector(0, 0);
isVisible = true;
}
public void Render() {
// If you forget to overshadow the Render() method you'll see this error message in your console
println("Error: A Drawable just defaulted to the catch-all Render(): '" + this.getClass() + "'.");
}
}
The Tile will be more sophisticated. It'll have more basic informations: row, column, is it currently selected (why not), a type like plains or hills or water, a bunch of neighboring tiles, a method to draw itself and a method to know if the unit can travel through it:
class Tile extends Drawable {
int row, column;
boolean selected = false;
TileType type;
ArrayList<Tile> neighbors = new ArrayList<Tile>();
Tile(int row, int column, TileType type) {
super(); // this calls the parent class' constructor
this.row = row;
this.column = column;
this.type = type;
// the hardcoded numbers are all cosmetics I included to make my grid looks less awful, nothing to see here
position.x = (_tileSize * 1.5) * (column + 1);
position.y = (_tileSize * 0.5) * (row + 1);
// this part checks if this is an offset row to adjust the spatial coordinates
if (row % 2 != 0) {
position.x += _tileSize * 0.75;
}
}
// this method looks recursive, but isn't. It doesn't call itself, but it calls it's twin from neighbors tiles
void FillCanTravelArrayList(int travelPoints, boolean originalTile) {
if (travelPoints >= type.travelCost) {
// if the unit has enough travel points, we add the tile to the "the unit can get there" list
if (!_canTravel.contains(this)) {
// well, only if it's not already in the list
_canTravel.add(this);
}
// then we check if the unit can go further
for (Tile t : neighbors) {
if (originalTile) {
t.FillCanTravelArrayList(travelPoints, false);
} else {
t.FillCanTravelArrayList(travelPoints - type.travelCost, false);
}
}
}
}
void Render() {
if (isVisible) {
// the type knows which colors to use, so we're letting the type draw the tile
type.Render(this);
}
}
}
The Tile Types
The TileType is a strange animal: it's a real class, but it's never used anywhere. That's because it's a common root for all tile types, which will inherit it's basics. The "City" tile may need very different variables than, say, the "Desert" tile. But both need to be able to draw themselves, and both need to be owned by the tiles.
class TileType {
// cosmetics
color fill = color(255, 255, 255);
color stroke = color(0);
float strokeWeight = 2;
// every tile has a "travelCost" variable, how much it cost to travel through it
int travelCost = 10;
// while I put this method here, it could have been contained in many other places
// I just though that it made sense here
void Render(Tile tile) {
fill(fill);
if (tile.selected) {
stroke(255);
} else {
stroke(stroke);
}
strokeWeight(strokeWeight);
DrawPolygon(tile.position.x, tile.position.y, _tileSize/2, 6);
textAlign(CENTER, CENTER);
fill(255);
text(tile.column + ", " + tile.row, tile.position.x, tile.position.y);
}
}
Each tile type can be custom, now, yet each tile is... just a tile, whatever it's content. Here are the TileType I used in this demonstration:
// each different tile type will adjust details like it's travel cost or fill color
class Plains extends TileType {
Plains() {
this.fill = color(0, 125, 0);
this.travelCost = 10;
}
}
class Water extends TileType {
// here I'm adding a random variable just to show that you can custom those types with whatever you need
int numberOfFishes = 10;
Water() {
this.fill = color(0, 0, 125);
this.travelCost = 1000;
}
}
class Hill extends TileType {
Hill() {
this.fill = color(125, 50, 50);
this.travelCost = 15;
}
}
Non-class methods
I added a mouseClicked() method so we can select a hex to check how far from it the unit can travel. In your game, you would have to make it so when you select a unit all these things fall into place, but as this is just a proof of concept the unit is imaginary and it's location is wherever you click.
void mouseClicked() {
// clearing the array which contains tiles where the unit can travel as we're changing those
_canTravel.clear();
for (Tile t : _tiles) {
// select the tile we're clicking on (and nothing else)
t.selected = IsPointInRadius(t.position, new PVector(mouseX, mouseY), _tileSize/2);
if (t.selected) {
// if a tile is selected, check how far the imaginary unit can travel
t.FillCanTravelArrayList(unitTravelPoints, true);
}
}
}
At last, I added 2 "helper methods" to make things easier:
// checks if a point is inside a circle's radius
boolean IsPointInRadius(PVector center, PVector point, float radius) {
// simple math, but with a twist: I'm not using the square root because it's costly
// we don't need to know the distance between the center and the point, so there's nothing lost here
return pow(center.x - point.x, 2) + pow(center.y - point.y, 2) <= pow(radius, 2);
}
// draw a polygon (I'm using it to draw hexagons, but any regular shape could be drawn)
void DrawPolygon(float x, float y, float radius, int npoints) {
float angle = TWO_PI / npoints;
beginShape();
for (float a = 0; a < TWO_PI; a += angle) {
float sx = x + cos(a) * radius;
float sy = y + sin(a) * radius;
vertex(sx, sy);
}
endShape(CLOSE);
}
How Travel is calculated
Behind the scenes, that's how the program knows where the unit can travel: in this example, the unit has 30 travel points. Plains cost 10, hills cost 15. If the unit has enough points left, the tile is marked as "can travel there". Every time a tile is in travel distance, we also check if the unit can get further from this tile, too.
Now, if you're still following me, you may ask: how do the tiles know which other tile is their neighbor? That's a great question. I suppose that an algorithm checking their coordinates would be the best way to handle this, but as this operation will happen only once when we create the map I decided to take the easy route and check which tiles were the close enough spatially:
void setup() {
// create the grid
for (int i=0; i<_gridWidth; i++) {
for (int j=0; j<_gridHeight; j++) {
int rand = (int)random(100);
if (rand < 20) {
_tiles.add(new Tile(j, i, new Water()));
} else if (rand < 50) {
_tiles.add(new Tile(j, i, new Hill()));
} else {
_tiles.add(new Tile(j, i, new Plains()));
}
}
}
// detect and save neighbor tiles for every Tile
for (Tile currentTile : _tiles) {
for (Tile t : _tiles) {
if (t != currentTile) {
if (IsPointInRadius(currentTile.position, t.position, _tileSize)) {
currentTile.neighbors.add(t);
}
}
}
}
}
Complete code for copy-pasting
Here's the whole thing in one place so you can easily copy and paste it into a Processing IDE and play around with it (also, it includes how I draw my awful grid):
// Debug
int unitTravelPoints = 30; // this is the number if "travel points" currently being tested, you can change it
// Golbals
float _tileSize = 60;
int _gridWidth = 10;
int _gridHeight = 20;
ArrayList<Tile> _tiles = new ArrayList<Tile>();
ArrayList<Tile> _canTravel = new ArrayList<Tile>();
void settings() {
// this is how to make a window size's dynamic
size((int)(((_gridWidth+1) * 1.5) * _tileSize), (int)(((_gridHeight+1) * 0.5) * _tileSize));
}
void setup() {
// create the grid
for (int i=0; i<_gridWidth; i++) {
for (int j=0; j<_gridHeight; j++) {
int rand = (int)random(100);
if (rand < 20) {
_tiles.add(new Tile(j, i, new Water()));
} else if (rand < 50) {
_tiles.add(new Tile(j, i, new Hill()));
} else {
_tiles.add(new Tile(j, i, new Plains()));
}
}
}
// detect and save neighbor tiles for every Tile
for (Tile currentTile : _tiles) {
for (Tile t : _tiles) {
if (t != currentTile) {
if (IsPointInRadius(currentTile.position, t.position, _tileSize)) {
currentTile.neighbors.add(t);
}
}
}
}
}
void draw() {
background(0);
// show the tiles
for (Tile t : _tiles) {
t.Render();
}
// show how far you can go
for (Tile t : _canTravel) {
fill(0, 0, 0, 0);
if (t.selected) {
stroke(255);
} else {
stroke(0, 255, 0);
}
strokeWeight(5);
DrawPolygon(t.position.x, t.position.y, _tileSize/2, 6);
}
}
class Drawable {
PVector position;
boolean isVisible;
public Drawable() {
position = new PVector(0, 0);
isVisible = true;
}
public void Render() {
// If you forget to overshadow the Render() method you'll see this error message in your console
println("Error: A Drawable just defaulted to the catch-all Render(): '" + this.getClass() + "'.");
}
}
class Tile extends Drawable {
int row, column;
boolean selected = false;
TileType type;
ArrayList<Tile> neighbors = new ArrayList<Tile>();
Tile(int row, int column, TileType type) {
super(); // this calls the parent class' constructor
this.row = row;
this.column = column;
this.type = type;
// the hardcoded numbers are all cosmetics I included to make my grid looks less awful, nothing to see here
position.x = (_tileSize * 1.5) * (column + 1);
position.y = (_tileSize * 0.5) * (row + 1);
// this part checks if this is an offset row to adjust the spatial coordinates
if (row % 2 != 0) {
position.x += _tileSize * 0.75;
}
}
// this method looks recursive, but isn't. It doesn't call itself, but it calls it's twin from neighbors tiles
void FillCanTravelArrayList(int travelPoints, boolean originalTile) {
if (travelPoints >= type.travelCost) {
// if the unit has enough travel points, we add the tile to the "the unit can get there" list
if (!_canTravel.contains(this)) {
// well, only if it's not already in the list
_canTravel.add(this);
}
// then we check if the unit can go further
for (Tile t : neighbors) {
if (originalTile) {
t.FillCanTravelArrayList(travelPoints, false);
} else {
t.FillCanTravelArrayList(travelPoints - type.travelCost, false);
}
}
}
}
void Render() {
if (isVisible) {
// the type knows which colors to use, so we're letting the type draw the tile
type.Render(this);
}
}
}
class TileType {
// cosmetics
color fill = color(255, 255, 255);
color stroke = color(0);
float strokeWeight = 2;
// every tile has a "travelCost" variable, how much it cost to travel through it
int travelCost = 10;
// while I put this method here, it could have been contained in many other places
// I just though that it made sense here
void Render(Tile tile) {
fill(fill);
if (tile.selected) {
stroke(255);
} else {
stroke(stroke);
}
strokeWeight(strokeWeight);
DrawPolygon(tile.position.x, tile.position.y, _tileSize/2, 6);
textAlign(CENTER, CENTER);
fill(255);
text(tile.column + ", " + tile.row, tile.position.x, tile.position.y);
}
}
// each different tile type will adjust details like it's travel cost or fill color
class Plains extends TileType {
Plains() {
this.fill = color(0, 125, 0);
this.travelCost = 10;
}
}
class Water extends TileType {
// here I'm adding a random variable just to show that you can custom those types with whatever you need
int numberOfFishes = 10;
Water() {
this.fill = color(0, 0, 125);
this.travelCost = 1000;
}
}
class Hill extends TileType {
Hill() {
this.fill = color(125, 50, 50);
this.travelCost = 15;
}
}
void mouseClicked() {
// clearing the array which contains tiles where the unit can travel as we're changing those
_canTravel.clear();
for (Tile t : _tiles) {
// select the tile we're clicking on (and nothing else)
t.selected = IsPointInRadius(t.position, new PVector(mouseX, mouseY), _tileSize/2);
if (t.selected) {
// if a tile is selected, check how far the imaginary unit can travel
t.FillCanTravelArrayList(unitTravelPoints, true);
}
}
}
// checks if a point is inside a circle's radius
boolean IsPointInRadius(PVector center, PVector point, float radius) {
// simple math, but with a twist: I'm not using the square root because it's costly
// we don't need to know the distance between the center and the point, so there's nothing lost here
return pow(center.x - point.x, 2) + pow(center.y - point.y, 2) <= pow(radius, 2);
}
// draw a polygon (I'm using it to draw hexagons, but any regular shape could be drawn)
void DrawPolygon(float x, float y, float radius, int npoints) {
float angle = TWO_PI / npoints;
beginShape();
for (float a = 0; a < TWO_PI; a += angle) {
float sx = x + cos(a) * radius;
float sy = y + sin(a) * radius;
vertex(sx, sy);
}
endShape(CLOSE);
}
Hope it'll help. Have fun!
You will have to use similar algorithms we use on pathfinding. you create a stack or queue that will hold a class storing the position of the hex and the number of moves left from that point, initially you insert your starting position with the number of moves you have and mark that hex as done ( to not re-use a position you have already been on ), then you pop an element, and you insert every neighbor of that hex with a number of moves -1. when you insert the hexes with zero moves, those are your endpoints. And before inserting any hex check if it's not already done.
I hope I was clear, your question was a bit vague but I tried to give you an idea of how these solutions are usually done, also I think your question fits more into algorithms rather then processing
Best of luck

libgdx pixel collision on very simple 2D game

I'm very new to libgdx and game dev. Ive been tasked with an assignment where I am given a very simple 2D game which has rectangular collision detection already implemented.
The game is just made up of a square which can be controlled by the player, which is inside of walls with other scattered squares within the walls. I now need to implement pixel perfect collision between the player square and the scattered squares/walls.
I originally tried using Pixmap and tried to check if the pixels of the current coordinate were transparent for the player rectangle and whichever rectangle it is colliding with but am having trouble implementing the code correctly.
I understand that I must just check the pixel colors for both objects at a certain coordinate and see if they both are not transparent for a collision to occur but I am having trouble. I have spent a very long time researching and looking online so any help or advice on how to execute this would be greatly appreciated!
public boolean entityCollision(Entity e1, Direction direction, float newX, float newY) {
boolean collision = false;
for(int i = 0; i < entities.size(); i++) {
Entity e2 = entities.get(i);
if (e1 != e2) {
if (newX < e2.x + e2.width && e2.x < newX + e1.width &&
newY < e2.y + e2.height && e2.y < newY + e1.height) {
collision = true;
int colorOne = player.getPixel((int)newX,(int)newY);
int colorTwo = enemy.getPixel((int)e2.x,(int)e2.y);
if (colorOne != 0 && colorTwo != 0) {
collision = true;
e1.entityCollision(e2, newX, newY, direction);
}
}
}
}
return collision;
}
Take a look at the Rectangle implementation of libgdx. It allows you to do something like this:
Rectangle e1Rec = new Rectangle(e1.x, e1.y, e1.width, e1.height);
Rectangle e2Rec = new Rectangle(e2.x, e2.y, e2.width, e2.height);
if (e1Rec.contains(e2Rec)) {
// Collision!
}
If you want collisions with a sprite, you get the sprite rectangle and check for collisions inside of that rectangle.
sprite.getboundingrectangle().contains(x, y) ; //returns a boolean

How to do collision detection with many walls (maze)?

In my game, the player navigates a maze. I can't figure out how to do proper collision detection with the walls. It is easy to do collision detection for staying in a certain area:
if (x > rightWallX - playerWidth) x = rightWallX - playerWidth;
if (x < leftWallX) x = leftWallX;
//...
But how would I do collision detection for many walls?
I can do plain collision detection without correction (like if (intersecting) return true;), but I can't correct this correctly. If I just store the old x and y and reset them, then
The object never actually touches the wall
If the object can go up but is blocked to the right, it won't go up, it will just not move.
How is collision detection in a maze done?
The easiest way, once you have solved collision detection, to fix the collision is to move the actor to the closest valid position to where the actor would be were it not for the object it collides with. This assumes no inertia, but it is sufficient for maze-like games or top-down map-crawling games.
If you want to simplify your calculations further, you can limit yourself to detecting if changing the actor's x or y coordinate would be better. If your actor has an axis-aligned rectangular hit-box and all obstacles are axis-aligned rectangular as well (the simplest case), this assumption is indeed correct. However, the results might not be satisfactory in some other cases (potential artifact: speed boost from gliding diagonal walls - not the case in most maze games).
Keep in mind multiple collisions could happen concurrently (pushing against two walls). If there are no sharp angles between two walls that an actor could both intersect (say, if all your obstacles are axis aligned and sufficiently spaced), fixing each collision in turn will suffice - just don't stop after the first collision.
You can use Rectangle.intersects() method:
public Rectangle Player(){
return new Rectangle(PlayerX,PlayerY,PlayerWidth,PlayerHeight);
//we do this for getting players x and y values every tick
}
if(Player().intersects(new Rectangle(0,0,100,50)))//if(player touching wall)
new Rectangle(0,0,100,50) is just an example you can change it.
Ok so i'm currently making a 2D top down view game and I'm not sure how you created your maze. However, in my game my Level is created from a Tile[][] tiles = new Tile[levelWidth][levelHeight]; array. The way i handled collision detection was by checking the surrounding tiles to see if they were solid.
This is my getTile method.
public Tile[][] getTile(int x, int y) {
if (x < 0 || x >= getWidth() || y < 0 || y >= getHeight()) {
return new VoidTile();
} else {
return tiles[x][y];
}
}
In my Tile.java class i have a isSolid() method which returns whether the tile is solid or not. All of my tiles extend my Tile.java so they inherit this method and I override it in their constructor. As i said previously, I am not sure whether or not you use the same style of level implementation as i do. However, It is good practice to do it this way :)
Personally, I am not a big fan of using the .intersects() and .contains() methods for Sprite collision detection. I mainly use them for buttons and alike.
Ok so,
In my player.java class i have a checkBlockedDirection(int x, int y) method and it looks like this.
public void checkBlockedDirection(int x, int y) {
boolean u = map.getTile(x, y - 1).isSolid();
boolean d = map.getTile(x, y + 1).isSolid();
boolean l = map.getTile(x - 1, y).isSolid();
boolean r = map.getTile(x + 1, y).isSolid();
if (u) {
uBlocked = true;
System.out.println("up tile blocked");
} else {
uBlocked = false;
}
if (d) {
dBlocked = true;
System.out.println("down tile blocked");
} else {
dBlocked = false;
}
if (l) {
lBlocked = true;
System.out.println("left tile blocked");
} else {
lBlocked = false;
}
if (r) {
rBlocked = true;
System.out.println("right tile blocked");
} else {
rBlocked = false;
}
}
Then in my player update method i have this
public void tick() {
float dx = 0;
float dy = 0;
if (input.up.isPressed()) {
direction = 0;
} else if (input.down.isPressed()) {
direction = 2;
} else if (input.left.isPressed()) {
direction = 3;
} else if (input.right.isPressed()) {
direction = 1;
} else {
direction = 4; // standing
}
checkBlockedDirection((int)x, (int)y);
if (input.up.isPressed() && y > 0 && !uBlocked) {
dy += -speed;
} else if (input.down.isPressed() && y < map.getHeight() - 1 && !dBlocked) {
dy += speed;
} else if (input.left.isPressed() && x > 0 && !lBlocked) {
dx += -speed;
} else if (input.right.isPressed() && x < map.getWidth() - 1 && !rBlocked) {
dx += speed;
}
x += dx;
y += dy;
}
Basically it just checks whether or not the blocks up, down, left, or right are solid. If they are solid then it wont move and if they arent solid then you can move in the desired direction.
Not sure if this helps or not but it's just my take on this kind of grid collision detection :)
Hope this helps :)
Enjoy

Java Tile Game - Collision Detection

I am making a java game using tiles. I am having trouble with the collision element. I am defining rectangles for each tile on the map, and another rectangle for the player. What I am having trouble with is knowing which side the player is coming from when it hits the rectangle, and then pushing the player off in the direction the player came from. I have already made a method that checks how much the character is inside the rectangle, so it can know how much to push it out, but I can't figure out how to tell which side the character is coming from.
here is my current collision methoe - note rect1 is the character and rect2 is the tile
public void collision(Rectangle rect1, Rectangle rect2) {
float xAdd;
float xAdd2;
float yAdd;
float yAdd2;
boolean hitRight = false;
boolean hitLeft = false;
boolean hitTop = false;
boolean hitBot = false;
Vector2f rect1Origin = new Vector2f(rect1.x, rect1.y);
Vector2f rect2Origin = new Vector2f(rect2.x, rect2.y);
Vector2f rect1Mid = new Vector2f((rect1.x + rect1.width) / 2,(rect1.y + rect1.height) / 2);
Vector2f rect2Mid = new Vector2f((rect2.x + rect2.width) / 2,(rect2.y + rect2.height) / 2);
Vector2f rect1A = new Vector2f(rect1Origin.x + rect1.width, rect1.y);
Vector2f rect1B = new Vector2f(rect1Origin.x, rect1Origin.y+ rect1.height);
Vector2f rect1C = new Vector2f(rect1Origin.x + rect1.width,rect1Origin.y + rect1.height);
Vector2f rect2A = new Vector2f(rect2Origin.x + rect2.width, rect2.y);
Vector2f rect2B = new Vector2f(rect2Origin.x, rect2Origin.y
+ rect2.height);
Vector2f rect2C = new Vector2f(rect2Origin.x + rect2.width,
rect2Origin.y + rect2.height);
xAdd = rect2C.x - rect1B.x;
xAdd2 = rect1C.x - rect2B.x;
yAdd = rect2A.y - rect1B.y;
yAdd2 = rect2C.y - rect1A.y;
if (rect1Mid.y < rect2Mid.y) {
if (rect1.intersects(rect2)) {
y_pos += yAdd;
}
}
if (rect1Mid.y > rect2Mid.y) {
if (rect1.intersects(rect2)) {
System.out.println(yAdd2);
y_pos += yAdd2;
}
}
if(rect1Mid.x > rect2Mid.x){
if(rect1.intersects(rect2)){
hitRight = true; x_pos += xAdd;
}
}
if(rect1Mid.x< rect2Mid.x){
if(rect1.intersects(rect2)) {
x_pos += -xAdd2;
}
}
}
Any help would be greatly appreciated
Thanks
Keep two positions for your character - where the it is (as of last frame, move, etc.), and where you want to move it. Then only move it if you don't detect collision - if you don't allow for corrupt state, you don't have to fix it.
Edit: The collision method should be boolean - it should be done before actually moving the character, like
if (!collision(character, tile))
{
doMove(character);
}
else
{
//custom handling if required
}
Edit2: The previous would work only for a small step, if you need a partial move, you really need to know the character's original position, like move(originalPosition, desiredPosition, tile), where you can deduce the direction from the originalPosition and the tile.
The main point is, you don't actually move the character, before you have a valid position for it.
First of all, any Sprite (the character and the tile here) should have four members: xPos, yPos, xVec, yVec. With that in mind, you know where the character was and will be in the next frame.
sprite.xPos += sprite.xVec;
sprite.yPos += sprite.yVec;
Also take a look at your code, you should improve the propositions (for instance, you check if(rect1.intersects(rect2)) in the four if statements). Instead, check only once for the intersects(), and inside check for each possible case (left, right, top, bottom hit).
Finally, the collision method should receive 2 sprites objects (for instance, sprite1 and sprite2), and then check for the intersects() method, taking into account the original positions of both sprites and its vectors. If the sprites intersect in the new position, then you invert the object's vector accordingly.
Assuming you can detect the collision with your code. Here is a way to determine the position of your character with respect to the rectangle object.
Find the middle point of the rectangle. ( Rx,Ry).
Find the middle point of your character sprite. ( Sx,Sy).
Now you can compare Rx,Ry , Sx , Sy to determine in which side is Sx and Sy of Rx and Ry
for ex:
if Sx < Rx then
Sx = left side of Rx
Was struggling with this myself, but after some thinking this check can be done by having all your entities in a List, in my code I had to block the player from moving through blocks. All blocks is in a List. Now, when checking players position I create a new Rectangle and put the players position one frame forward in the direction the update will happened next frame, if the the rectangle intersects that update for that direction will not happened. Heres my code for this:
if (isKeyPressed(KeyEvent.VK_LEFT)) {
if(!collisionWithBlocks(1)){
pl.x = pl.x - updatePlayerPosition;
}
}
if (isKeyPressed(KeyEvent.VK_RIGHT)) {
if(!collisionWithBlocks(0)){
pl.x = pl.x + updatePlayerPosition;
}
}
if (isKeyPressed(KeyEvent.VK_UP)) {
if(!collisionWithBlocks(3)){
pl.y = pl.y - updatePlayerPosition;
}
}
if (isKeyPressed(KeyEvent.VK_DOWN)) {
if(!collisionWithBlocks(2)){
pl.y = pl.y + updatePlayerPosition;
}
}
collisionWithBlocks():
public boolean collisionWithBlocks(int side){
for(Block b : main.blocks){
Rectangle block = b.getBounds();
Rectangle player = null;
if(side == 0){
player = new Rectangle(pl.x + updatePlayerPosition, pl.y, pl.getWidth(), pl.getHeight());
} else if(side == 1){
player = new Rectangle(pl.x - updatePlayerPosition, pl.y, pl.getWidth(), pl.getHeight());
} else if(side == 2){
player = new Rectangle(pl.x, pl.y + updatePlayerPosition, pl.getWidth(), pl.getHeight());
} else if(side == 3){
player = new Rectangle(pl.x, pl.y - updatePlayerPosition, pl.getWidth(), pl.getHeight());
}
if(player.intersects(block)){
return true;
}
}
return false;
}
updatePlayerPosition is 2, and changes in my code but enough about that.
In your case, I recommend putting the entities in a List and then checking like I did over.

Constructing a lists with same elements throughout in Java

class Board {
Direction[][] board;
Integer width;
Integer height;
Board(Integer width, Integer height) {
this.width = width;
this.height = height;
board = new Direction[width][height];
for (Integer y = 0; y < height; y++) {
for (Integer x = 0; x < width; x++) {
board[x][y] = Direction.Down;
}
}
}
Direction direction(Integer x, Integer y) {
return board[x][y];
}
void update(Integer x, Integer y) {
board[x][y].next();
}
}
The enumerated values in Direction are: Up or Down
Firstly, I want to initialise my board so that all the initial values at every place of the board are set to Down.
Secondly, I want to retrieve the value at a particular (x,y) position of of the board
Thirdly, I want to be able to update the value at that particular element of the board. The next() function is designed to change the value from Down to Up or vice versa. This is declared separately.
I believe I am doing something wrong as the code compiles, however my program doesn't work as it should.
If the problem is your update method:
Values of an enum type should be immutable. Calling the next() method on it could change the enum object (depending on its implementation), but as you are using the same object in all of your array positions, all of those would change, too.
You might want to use something like this:
board[x][y] = board[x][y].next();
with
enum Direction {
UP, DOWN;
public Direction next() {
if (this == UP)
return DOWN;
else
return UP;
}
}
Most likely, you want to change your update code to be:
board[x][y] = board[x][y].next();
Of course, I'm assuming a lot about what's not working with your code, but it's a common mistake I've seen before.

Categories

Resources