I'm making a program that deals with a large collection of 2D data. I'm displaying this data in an XYPlot. Is there some way I can mark which direction the line is moving over time? The data bounces all over the plot, but needs to be in a specific order, so a scatter plot won't work for me.
EDIT: To clarify - Neither the X or Y axis is time. This is a parametric plot.
I have discovered a possible solution. It needs some working through to make it look right, but it works for now.
There is a class in JFreeChart called XYPointerAnnotation, which allows you to add arrows to your chart of a custom angle, length, and width. I'm still trying to figure the class out, but so far it seems to do well enough for my purposes. Code attached:
for (int i = 0;i < data.getItemCount(0) - 1;i++)
{
double x1 = data.getSeries(0).getDataItem(i).getXValue();
double x2 = data.getSeries(0).getDataItem(i + 1).getXValue();
double y1 = data.getSeries(0).getDataItem(i).getYValue();
double y2 = data.getSeries(0).getDataItem(i + 1).getYValue();
double angle = Math.atan2(y1 - y2, x2 - x1) + Math.PI;
XYPointerAnnotation arrow = new XYPointerAnnotation("",x1,y1,angle);
if (i == 0)
{
arrow.setText("Start");
}
else if (i % 5 == 0)
{
arrow.setText(Integer.toString(i));
}
arrow.setLabelOffset(15.0);
arrow.setToolTipText(Integer.toString(i));
_originalPlot.addAnnotation(arrow);
}
Related
I am trying to interpolate color along a line so that, given two points and their respective RGB values, I can draw a line with a smooth color gradient. Using Bresenham's Line Algorithm, I can now draw lines, but am not sure how to begin interpolating colors between the two end points. The following is part of the drawLine() function that works for all line whose slope are less than 1.
int x_start = p1.x, x_end = p2.x, y_start =p1.y, y_end = p2.y;
int dx = Math.abs(x_end-x_start), dy = Math.abs(y_end-y_start);
int x = x_start, y = y_start;
int step_x = x_start < x_end ? 1:-1;
int step_y = y_start < y_end ? 1:-1;
int rStart = (int)(255.0f * p1.c.r), rEnd = (int)(255.0f * p2.c.r);
int gStart = (int)(255.0f * p1.c.g), gEnd = (int)(255.0f * p2.c.g);
int bStart = (int)(255.0f * p1.c.b), bEnd = (int)(255.0f * p2.c.b);
int xCount = 0;
//for slope < 1
int p = 2*dy-dx;
int twoDy = 2*dy, twoDyMinusDx = 2*(dy-dx);
int xCount = 0;
// draw the first point
Point2D start = new Point2D(x, y, new ColorType(p1.c.r, p1.c.g, p1.c.b));
drawPoint(buff, start);
float pColor = xCount / Math.abs((x_end - x_start));
System.out.println(x_end + " " + x_start);
while(x != x_end){
x+= step_x;
xCount++;
if(p<0){
p+= twoDy;
}
else{
y += step_y;
p += twoDyMinusDx;
}
Point2D draw_line = new Point2D(x, y, new ColorType(p1.c.r*(1-pColor)+p2.c.r*pColor,p1.c.g*(1-pColor)+p2.c.g*pColor,p1.c.b*(1-pColor)+p2.c.b*pColor));
System.out.println(pColor);
drawPoint(buff,draw_line );
}
So what I'm thinking is that, just like drawing lines, I also need some sort of decision parameter p to determine when to change the RGB values. I am thinking of something along lines of as x increments, look at each rgb value and decide if I want to manipualte them or not.
I initialized rStart and rEnd(and so on for g and b) but have no idea where to start. any kind of help or suggestions would be greatly appreciated!
Edit: thanks #Compass for the great suggestion ! Now I've ran into another while trying to implementing that strategy, and I am almost certain it's an easy bug. I just can't see it right now. For some reason my pColor always return 0, I am not sure why. I ran some print statements to make sure xCount is indeed increasing, so I am not sure what else might've made this variable always 0.
I remember figuring this out way back when I was learning GUI! I'll explain the basic concepts for you.
Let's say we have two colors,
RGB(A,B,C)
and
RGB(X,Y,Z)
for simplicity.
If we know the position percentage-wise (we'll call this P, a float 0 for beginning, 1.0 at end) along the line, we can calculate what color should be there using the following:
Resultant Color = RGB(A*(1-P)+X*P,B*(1-P)+Y*P,C*(1-P)+Z*P)
In other words, you average out the individual RGB values along the line.
Actually you will be drawing the line in RGB space as well !
Bresenham lets you compute point coordinates from (X0, Y0) to (X1, Y1).
This is done by a loop on X or Y, with a linear interpolation on the other coordinate.
Just extend the algorithm to draw a line from (X0, Y0, R0, G0, B0) to (X1, Y1, R1, G1, B1), in the same loop on X or Y, with a linear interpolation on the other coordinates.
I have spent hours looking for the solution to this: I am developing a little top-down game with libgdx (maybe it matters what engine i am using). Now i have to implement the collision detection between my character (circle) and the wall (rectangle). I want the character to slide along the wall on collision, if sliding is possible.
Let me explain:
If i am moving 45 degrees right up i can collide with the down, the
left or the corner of a wall.
If i collide with the left i want to stop x-movement and move only up. If i leave the wall then i want to go on moving right up. The same
with the down side (stop y-movement)
If i collide with the Corner i want to stop movement (sliding not possible).
What i am doing actually is to check if the left line of the rectangle intersects my circle. Then i check intersection between the left line of wall and my circle and the bottom line of wall and my circle. Depending on which intersection occuret i set back x/y possition of my circle and set x/y Speed to 0. The Problem is, that most times not a collision bt an overlap occures. So the bottom check returns true, even if in reality the circle would only collide with the right. In this case both intersection test would return true and i would reset both speeds like on the Corner collision.
How can i solve this Problem? Is ther a better way to detect collision and collision side or corner?
I don't Need the exact Point of collision just the side of the rectangle.
Edit:
I have to say, that the rects aren't rotated just parallel to the x-axis.
You can find an explanation for circle/rectangle collision below, but please note that this type of collision might not be necessary for your needs. If, for example, you had a rectangle bounding box for your character the algorithm would be simpler and faster. Even if you are using a circle, it is probable that there is a simpler approach that is good enough for your purposes.
I though about writing the code for this, but it would take too long so here is only an explanation:
Here is a example movement of your character circle, with its last (previous) and current positions. Wall rectangle is displayed above it.
Here is that same movement, dotted lines represent the area the circle sweeps in this move. The sweep area is capsule shaped.
It would be difficult to calculate the collision of these two object, so we need to do this differently. If you look at the capsule on the previous image, you will see that it is simply the movement line extended in every direction by the radius of the circle. We can move that "extension" from the movement line to the wall rectangle. This way we get a rounded rectangle like on the image below.
The movement line will collide with this extended (rounded) rectangle if and only if the capsule collides with the wall rectangle, so they are somehow equivalent and interchangeable.
Since this collision calculation is still non-trivial and relatively expensive, you can first do a fast collision check between the extended wall rectangle (non-rounded this time) and the bounding rectangle of the movement line. You can see these rectangles on the image below - they are both dotted. This is a fast and easy calculation, and while you play the game there will probably NOT be an overlap with a specific wall rectangle >99% of the time and collision calculation will stop here.
If however there is an overlap, there is probably a collision of the character circle with wall rectangle, but it is not certain as will be demonstrated later.
Now you need to calculate the intersection between the movement line itself (not its bounding box) and the extended wall rectangle. You can probably find an algorithm how to do this online, search for line/rectangle intersection, or line/aabb intersection (aabb = Axis Aligned Bounding Box). The rectangle is axis-aligned and this makes the calculation simpler. The algorithm can give you intersection point or points since it is possible that there are two - in this case you choose the closest one to the starting point of the line. Below is an example of this intersection/collision.
When you get an intersection point, it should be easy to calculate on which part of the extended rectangle this intersection is located. You can see these parts on the image above, separated by red lines and marked with one or two letters (l - left, r - right, b - bottom, t - top, tl - top and left etc).
If the intersection is on parts l, r, b or t (the single letter ones, in the middle) then you are done. There is definitely a collision between character circle and wall rectangle, and you know on which side. In the example above, it is on the bottom side. You should probably use 4 variables called something like isLeftCollision, isRightCollision, isBottomCollsion and isTopCollision. In this case you would set isBottomCollision to true, while the other 3 would remain at false.
However, if the intersection is on the corner, on the two-letter sections, additional calculations are needed to determine if there is an actual collision between character circle and wall rectangle. Image below shows 3 such intersections on the corners, but there is an actual circle-rectangle collision on only 2 of them.
To determine if there is a collision, you need to find an intersection between the movement line and the circle centered in the closest corner of the original non-extended wall rectangle. The radius of this circle is equal to the radius of character circle. Again, you can google for line/circle intersection algorithm (maybe even libgdx has one), it isn't complex and shouldn't be hard to find.
There is no line/circle intersection (and no circle/rectangle collision) on bl part, and there are intersections/collisions on br and tr parts.
In the br case you set both isRightCollision, isBottomCollsion to true and in the tr case you set both isRightCollision and isTopCollision to true.
There is also one edge case you need to look out for, and you can see it on the image below.
This can happen if the movement of previous step ends in the corner of the the extended rectangle, but outside the radius of the inner rectangle corner (there was no collision).
To determine if this is the case, simply check if movement staring point is inside the extended rectangle.
If it is, after the initial rectangle overlap test (between extended wall rectangle and bounding rectangle of movement line), you should skip line/rectangle intersection test (because in this case there might not be any intersection AND still be a circle/rectangle collision), and also simply based on movement stating point determine which corner you are in, and then only check for line/circle intersection with that corner's circle. If there is intersection, there is a character circle/wall rectangle collision, otherwise not.
After all of this, the collision code should be simple:
// x, y - character coordinates
// r - character circle radius
// speedX, speedY - character speed
// intersectionX, intersectionY - intersection coordinates
// left, right, bottom, top - wall rect positions
// I strongly recomment using a const "EPSILON" value
// set it to something like 1e-5 or 1e-4
// floats can be tricky and you could find yourself on the inside of the wall
// or something similar if you don't use it :)
if (isLeftCollision) {
x = intersectionX - EPSILON;
if (speedX > 0) {
speedX = 0;
}
} else if (isRightCollision) {
x = intersectionX + EPSILON;
if (speedX < 0) {
speedX = 0;
}
}
if (isBottomCollision) {
y = intersectionY - EPSILON;
if (speedY > 0) {
speedY = 0;
}
} else if (isTopCollision) {
y = intersectionY + EPSILON;
if (speedY < 0) {
speedY = 0;
}
}
[Update]
Here is a simple and I believe efficient implementation of segment-aabb intersection that should be good enough for your purposes. It is a slightly modified Cohen-Sutherland algorithm. Also you can check out the second part of this answer.
public final class SegmentAabbIntersector {
private static final int INSIDE = 0x0000;
private static final int LEFT = 0x0001;
private static final int RIGHT = 0x0010;
private static final int BOTTOM = 0x0100;
private static final int TOP = 0x1000;
// Cohen–Sutherland clipping algorithm (adjusted for our needs)
public static boolean cohenSutherlandIntersection(float x1, float y1, float x2, float y2, Rectangle r, Vector2 intersection) {
int regionCode1 = calculateRegionCode(x1, y1, r);
int regionCode2 = calculateRegionCode(x2, y2, r);
float xMin = r.x;
float xMax = r.x + r.width;
float yMin = r.y;
float yMax = r.y + r.height;
while (true) {
if (regionCode1 == INSIDE) {
intersection.x = x1;
intersection.y = y1;
return true;
} else if ((regionCode1 & regionCode2) != 0) {
return false;
} else {
float x = 0.0f;
float y = 0.0f;
if ((regionCode1 & TOP) != 0) {
x = x1 + (x2 - x1) / (y2 - y1) * (yMax - y1);
y = yMax;
} else if ((regionCode1 & BOTTOM) != 0) {
x = x1 + (x2 - x1) / (y2 - y1) * (yMin - y1);
y = yMin;
} else if ((regionCode1 & RIGHT) != 0) {
y = y1 + (y2 - y1) / (x2 - x1) * (xMax - x1);
x = xMax;
} else if ((regionCode1 & LEFT) != 0) {
y = y1 + (y2 - y1) / (x2 - x1) * (xMin - x1);
x = xMin;
}
x1 = x;
y1 = y;
regionCode1 = calculateRegionCode(x1, y1, r);
}
}
}
private static int calculateRegionCode(double x, double y, Rectangle r) {
int code = INSIDE;
if (x < r.x) {
code |= LEFT;
} else if (x > r.x + r.width) {
code |= RIGHT;
}
if (y < r.y) {
code |= BOTTOM;
} else if (y > r.y + r.height) {
code |= TOP;
}
return code;
}
}
Here is some code example usage:
public final class Program {
public static void main(String[] args) {
float radius = 5.0f;
float x1 = -10.0f;
float y1 = -10.0f;
float x2 = 31.0f;
float y2 = 13.0f;
Rectangle r = new Rectangle(3.0f, 3.0f, 20.0f, 10.0f);
Rectangle expandedR = new Rectangle(r.x - radius, r.y - radius, r.width + 2.0f * radius, r.height + 2.0f * radius);
Vector2 intersection = new Vector2();
boolean isIntersection = SegmentAabbIntersector.cohenSutherlandIntersection(x1, y1, x2, y2, expandedR, intersection);
if (isIntersection) {
boolean isLeft = intersection.x < r.x;
boolean isRight = intersection.x > r.x + r.width;
boolean isBottom = intersection.y < r.y;
boolean isTop = intersection.y > r.y + r.height;
String message = String.format("Intersection point: %s; isLeft: %b; isRight: %b; isBottom: %b, isTop: %b",
intersection, isLeft, isRight, isBottom, isTop);
System.out.println(message);
}
long startTime = System.nanoTime();
int numCalls = 10000000;
for (int i = 0; i < numCalls; i++) {
SegmentAabbIntersector.cohenSutherlandIntersection(x1, y1, x2, y2, expandedR, intersection);
}
long endTime = System.nanoTime();
double durationMs = (endTime - startTime) / 1e6;
System.out.println(String.format("Duration of %d calls: %f ms", numCalls, durationMs));
}
}
This is the result I get from executing this:
Intersection point: [4.26087:-2.0]; isLeft: false; isRight: false; isBottom: true, isTop: false
Duration of 10000000 calls: 279,932343 ms
Please note that this is desktop performance, on an i5-2400 CPU. It will probably be much slower on Android devices, but I believe still more than sufficient.
I only tested this superficially, so if you find any errors, let me know.
If you use this algorithm, I believe you don't need special handling for that case where starting point is in the corner of the extended wall rectangle, since in this case you will get the intersection point at line start, and the collision detection procedure will continue to the next step (line-circle collision).
I suppose you determine the collision by calculating the distance of the circles center with the lines.
We can simplify the case and tell that the circle colliding with the corner if both distances are equal and smaller than the radius. The equality should have a tolerance of course.
More - may be not necessary- realistic approach would be to consider x,y speed and factor it in the equality check.
I have given a diagram of my current small problem that I need help with. My main purpose is to keep the point from going outside the circle. Nothing else.
The center of the circle is positioned at (x, y).
I only solved a little bit of the problem, and that is the collision detection part of my problem, as given below:
public void bound(Point p, Circle c){
double distance = Math.hypot(p.x - c.x, p.y - c.y);
if (distance >= c.radius){
//Clueless from here on out.
}
}
The part where I left a comment is the spot I couldn't figure anything out. I did tried to set the point's velocityX and velocityY to 0, but I realized the point will just stay put whenever it touches the circle.
So, I'm sort of stuck.
I have resolved this issue.
public void reflect(Hole h){
//R = -2*(V dot N)*N + V
//N is normalized.
double nx = (this.position[0]+this.diameter/2) - (h.x+16);
double ny = (this.position[1]+this.diameter/2) - (h.y+16);
double nd = Math.hypot(nx, ny);
if (nd == 0)
nd = 1;
nx /= nd;
ny /= nd;
double dotProduct = this.speed[0]*nx+this.speed[1]*ny;
this.speed[0] += (float)(-2*dotProduct*nx);
this.speed[1] += (float)(-2*dotProduct*ny);
}
public void reflectResponse() {
for (int i = 0; i <= 1; i++) {
position[i] -= speed[i];
speed[i] *= 0.992f;
}
}
I tried Oli Charlesworth's method from the comments, but it made things more... "complicated" than I expected. Someone else mentioned I used a completely 100%, vector-based algorithm, since I'm relying a lot on vector-based movements.
TIPS TO THOSE WHO DO READ THIS:
If you're working on object movements and collisions with vectors, seek vector-based algorithms.
If you're working with angles (either degrees or radians), use Oli Charlesworth's method.
I need to draw a smooth line through a set of vertices. The set of vertices is compiled by a user dragging their finger across a touch screen, the set tends to be fairly large and the distance between the vertices is fairly small. However, if I simply connect each vertex with a straight line, the result is very rough (not-smooth).
I found solutions to this which use spline interpolation (and/or other things I don't understand) to smooth the line by adding a bunch of additional vertices. These work nicely, but because the list of vertices is already fairly large, increasing it by 10x or so has significant performance implications.
It seems like the smoothing should be accomplishable by using Bezier curves without adding additional vertices.
Below is some code based on the solution here:
http://www.antigrain.com/research/bezier_interpolation/
It works well when the distance between the vertices is large, but doesn't work very well when the vertices are close together.
Any suggestions for a better way to draw a smooth curve through a large set of vertices, without adding additional vertices?
Vector<PointF> gesture;
protected void onDraw(Canvas canvas)
{
if(gesture.size() > 4 )
{
Path gesturePath = new Path();
gesturePath.moveTo(gesture.get(0).x, gesture.get(0).y);
gesturePath.lineTo(gesture.get(1).x, gesture.get(1).y);
for (int i = 2; i < gesture.size() - 1; i++)
{
float[] ctrl = getControlPoint(gesture.get(i), gesture.get(i - 1), gesture.get(i), gesture.get(i + 1));
gesturePath.cubicTo(ctrl[0], ctrl[1], ctrl[2], ctrl[3], gesture.get(i).x, gesture.get(i).y);
}
gesturePath.lineTo(gesture.get(gesture.size() - 1).x, gesture.get(gesture.size() - 1).y);
canvas.drawPath(gesturePath, mPaint);
}
}
}
private float[] getControlPoint(PointF p0, PointF p1, PointF p2, PointF p3)
{
float x0 = p0.x;
float x1 = p1.x;
float x2 = p2.x;
float x3 = p3.x;
float y0 = p0.y;
float y1 = p1.y;
float y2 = p2.y;
float y3 = p3.y;
double xc1 = (x0 + x1) / 2.0;
double yc1 = (y0 + y1) / 2.0;
double xc2 = (x1 + x2) / 2.0;
double yc2 = (y1 + y2) / 2.0;
double xc3 = (x2 + x3) / 2.0;
double yc3 = (y2 + y3) / 2.0;
double len1 = Math.sqrt((x1-x0) * (x1-x0) + (y1-y0) * (y1-y0));
double len2 = Math.sqrt((x2-x1) * (x2-x1) + (y2-y1) * (y2-y1));
double len3 = Math.sqrt((x3-x2) * (x3-x2) + (y3-y2) * (y3-y2));
double k1 = len1 / (len1 + len2);
double k2 = len2 / (len2 + len3);
double xm1 = xc1 + (xc2 - xc1) * k1;
double ym1 = yc1 + (yc2 - yc1) * k1;
double xm2 = xc2 + (xc3 - xc2) * k2;
double ym2 = yc2 + (yc3 - yc2) * k2;
// Resulting control points. Here smooth_value is mentioned
// above coefficient K whose value should be in range [0...1].
double k = .1;
float ctrl1_x = (float) (xm1 + (xc2 - xm1) * k + x1 - xm1);
float ctrl1_y = (float) (ym1 + (yc2 - ym1) * k + y1 - ym1);
float ctrl2_x = (float) (xm2 + (xc2 - xm2) * k + x2 - xm2);
float ctrl2_y = (float) (ym2 + (yc2 - ym2) * k + y2 - ym2);
return new float[]{ctrl1_x, ctrl1_y, ctrl2_x, ctrl2_y};
}
Bezier Curves are not designed to go through the provided points! They are designed to shape a smooth curve influenced by the control points.
Further you don't want to have your smooth curve going through all data points!
Instead of interpolating you should consider filtering your data set:
Filtering
For that case you need a sequence of your data, as array of points, in the order the finger has drawn the gesture:
You should look in wiki for "sliding average".
You should use a small averaging window. (try 5 - 10 points). This works as follows: (look for wiki for a more detailed description)
I use here an average window of 10 points:
start by calculation of the average of points 0 - 9, and output the result as result point 0
then calculate the average of point 1 - 10 and output, result 1
And so on.
to calculate the average between N points:
avgX = (x0+ x1 .... xn) / N
avgY = (y0+ y1 .... yn) / N
Finally you connect the resulting points with lines.
If you still need to interpolate between missing points, you should then use piece - wise cubic splines.
One cubic spline goes through all 3 provided points.
You would need to calculate a series of them.
But first try the sliding average. This is very easy.
Nice question. Your (wrong) result is obvious, but you can try to apply it to a much smaller dataset, maybe by replacing groups of close points with an average point. The appropriate distance in this case to tell if two or more points belong to the same group may be expressed in time, not space, so you'll need to store the whole touch event (x, y and timestamp). I was thinking of this because I need a way to let users draw geometric primitives (rectangles, lines and simple curves) by touch
What is this for? Why do you need to be so accurate? I would assume you only need something around 4 vertices stored for every inch the user drags his finger. With that in mind:
Try using one vertex out of every X to actually draw between, with the middle vertex used for specifying the weighted point of the curve.
int interval = 10; //how many points to skip
gesture.moveTo(gesture.get(0).x, gesture.get(0).y);
for(int i =0; i +interval/2 < gesture.size(); i+=interval)
{
Gesture ngp = gesture.get(i+interval/2);
gesturePath.quadTo(ngp.x,ngp.y, gp.x,gp.y);
}
You'll need to adjust this to actually work but the idea is there.
I am making a small game, 2D, and I have a player.
EDIT:
This is what I have right now:
int oldX = player.x;
int oldY = player.y;
int newX = oldX - player.getSpeedX();
int newY = oldY - player.getSpeedY();
if(player.getBounds().intersects(brush1.getBounds())){
player.x = newX;
player.y = newY;
}else{
player.x = oldX;
player.y = oldY;
}
But, it is acting really weird, it changes speed when I go in from one side, etc.
For a formula and code that checks the intersection of a line segment and a circle, have a look at this SO answer.
The rest of the code should be quite clear, before you make a move, check if a collision occurs, if it would, don't move.
Depending on the behaviour you prefer, you could also move as close to the wall as possible and then move parallel to it to let the circle "slide" along the wall. This can be done by projecting the movement vector on a line with the same direction as the wall.
EDIT: Some comments on how to use the code from the answer to check for collisions:
The code uses a Dot function that computes the dot product of two vectors. You can either create a Vector class (a good exercise and it is useful for such projects) or compute just the dot product directly using the formulas here.
A vector class will make some of the code easier to read, but you can also operate on floats directly. For example, let's have a look at the computation of d (Direction vector of ray) and f (Vector from center sphere to ray start) as described at the start of the answer. With a vector class, this will be as easy as the formulas used there:
// L, E, C, d and f are of some vector type
d = L - E;
f = E - C;
Without a vector class, you'll have seperate x/y coordinates for all of the variables, bloating the code a bit but doing just the same:
// Twice as much variables, but all of them are usual floats.
dx = Lx - Ex;
dy = Ly - Ey;
fx = Ex - Cx;
fy = Ey - Cy;
I think your code snippet has a few bugs. Suggested fixes:
int oldX = player.x;
int oldY = player.y;
// *add* speed to old pos to get new pos
int newX = oldX + player.getSpeedX();
int newY = oldY + player.getSpeedY();
if(player.getBounds().intersects(brush1.getBounds())){
// Collision! Keep old position.
// Reverse speed (just a suggestion).
player.setSpeedX(-player.getSpeedX());
player.setSpeedY(-player.getSpeedY());
}else{
// No collision. Set position to *new* pos, not old pos.
player.x = newX;
player.y = newY;
}