Develop Reference

Kinect for Processing error when using multiple cameras

I'm trying to run two Kinect V1 cameras simultaneously in Processing 3. I have gotten to a solution that is not sustainable, and am I trying to make something more stable/reliable.
At the moment, whenever I try to run both cameras simultaneously on a single sketch, I am hit with the error
"Could not claim interface on camera: -3
Failed to open camera subdevice or it is not disabled.Failed to open motor subddevice or it is not disabled.Failed to open audio subdevice or it is not disabled.There are no kinects, returning null"
One camera opens, the other does not. It is not always consistent which camera opens, which leads me to believe there's something tripping over permissions after the objects are created, or when the second object is initialized.
My code is as such
import SimpleOpenNI.*;
import org.openkinect.freenect.*;
import org.openkinect.freenect2.*;
import org.openkinect.processing.*;
import org.openkinect.tests.*;
//Imported libraries
//Some might be unnecessary but I don't have time to check
//Better safe than sorry, maybe I'll delete later
Kinect kinect;
Kinect kinect2;
PImage depthImage;
PImage depthImage2;
//Set depth threshold
float minDepth = 996;
float maxDepth = 2493;
float iWidth1 = 0;
float iHeight1 = 0;
float iWidth2 = 0;
float iHeight2 = 0;
//Double check for the number of devices, mostly for troubleshooting
int numDevices = 0;
//control which device is being controlled (in case I want device control)
int deviceIndex = 0;
void setup() {
//set Arbitrary size
size(640, 360);
//Set up window to resize, need to figure out how to keep things centered
surface.setResizable(true);
//not necessary, but good for window management. Window label
surface.setTitle("KINECT 1");
//get number of devices, print to console
numDevices = Kinect.countDevices();
println("number of V1 Kinects = "+numDevices);
//set up depth for the first kinect tracking
kinect = new Kinect(this);
kinect.initDepth();
//Blank Image
depthImage = new PImage(kinect.width, kinect.height);
//set up second window
String [] args = {"2 Frame Test"};
SecondApplet sa = new SecondApplet();
PApplet.runSketch(args, sa);
}
//Draw first window's Kinect Threshold
void draw () {
if ((width/1.7778) < height) {
iWidth1 = width;
iHeight1 = width/1.7778;
} else {
iWidth1 = height*1.7778;
iHeight1 = height;
}
//Raw Image
image(kinect.getDepthImage(), 0, 0, iWidth1, iHeight1);
//Threshold Equation
int[] rawDepth = kinect.getRawDepth();
for (int i=0; i < rawDepth.length; i++) {
if (rawDepth[i] >= minDepth && rawDepth[i] <= maxDepth) {
depthImage.pixels[i] = color(255);
} else {
depthImage.pixels[i] = color(1);
}
}
}
public class SecondApplet extends PApplet {
public void settings() {
//arbitrary size
size(640, 360);
kinect2 = new Kinect(this);
kinect2.initDepth();
//Blank Image
depthImage2 = new PImage(kinect2.width, kinect2.height);
}
void draw () {
if ((width/1.7778) < height) {
iWidth2 = width;
iHeight2 = width/1.7778;
} else {
iWidth2 = height*1.7778;
iHeight2 = height;
}
image(kinect2.getDepthImage(), 0, 0, iWidth2, iHeight2);
surface.setResizable(true);
surface.setTitle("KINECT 2");
int[] rawDepth2 = kinect2.getRawDepth();
for (int i=0; i < rawDepth2.length; i++) {
if (rawDepth2[i] >= minDepth && rawDepth2[i] <= maxDepth) {
depthImage2.pixels[i] = color(255);
} else {
depthImage2.pixels[i] = color(1);
}
}
}
}
Curiously, the code returns a confirmation that there are two kinect devices connected in the console. For some reason, it cannot access both at the same time.
I'm not a very experienced code, so this code might look amateur. Open to feedback on other parts, but really just looking to solve this problem.
This code returns the error pasted above when there are two Kinect V1's connected to the computer.
Running Mac OS11.6.8 on an Intel MacBook Pro
Using Daniel Schiffman's OpenKinect for Processing as a starting point for the code.
I've run a successful iteration of this code with a slimmed down version of Daniel Schiffman's Depth Threshold example.
import org.openkinect.freenect.*;
import org.openkinect.processing.*;
Kinect kinect;
// Depth image
PImage depthImg;
// Which pixels do we care about?
// These thresholds can also be found with a variaty of methods
float minDepth = 996;
float maxDepth = 2493;
// What is the kinect's angle
float angle;
void setup() {
size(1280, 480);
kinect = new Kinect(this);
kinect.initDepth();
angle = kinect.getTilt();
// Blank image
depthImg = new PImage(kinect.width, kinect.height);
}
void draw() {
// Draw the raw image
image(kinect.getDepthImage(), 0, 0);
// Calibration
//minDepth = map(mouseX,0,width, 0, 4500);
//maxDepth = map(mouseY,0,height, 0, 4500);
// Threshold the depth image
int[] rawDepth = kinect.getRawDepth();
for (int i=0; i < rawDepth.length; i++) {
if (rawDepth[i] >= minDepth && rawDepth[i] <= maxDepth) {
depthImg.pixels[i] = color(255);
} else {
depthImg.pixels[i] = color(0);
}
}
// Draw the thresholded image
depthImg.updatePixels();
image(depthImg, kinect.width, 0);
//Comment for Calibration
fill(0);
text("TILT: " + angle, 10, 20);
text("THRESHOLD: [" + minDepth + ", " + maxDepth + "]", 10, 36);
//Calibration Text
//fill(255);
//textSize(32);
//text(minDepth + " " + maxDepth, 10, 64);
}
Using this code, I was able to get both cameras operating using the following process:
Connect a single Kinect v1 to the computer
Open and run the above code
Duplicate the sketch file
Connect the second Kinect V1 to the computer
Open and run the duplicated sketch of the same code
This worked for my purposes and remained stable for an extended period of time. However, this isn't a sustainable solution if anyone other than me wants to utilize this program.
Any help with this problem would be greatly appreciated

Camera doesnt rotate x axis using javafx

ive been using javafx recently and while i was doing so ive caught a problem in my code but spending a lot of time on it (mostly googling it) i cannot find the solution to my problem.
The problem happens when i try to rotate the camera using the x axis but instead of doing what i was wanting to happen instead it rotates around (0,0,0). Im not sure if it does rotate around (0,0,0) but that was the solution i could figure out. My camera starts staring at a cube but when tun left (adds 2 to the x axis) and the box goes a big circle. When turning right the box moves the other way around the circle. When applying the axis to the box, the box goes around fine.
My code is a bit messy but what i tried is to get the movement of turning and the box should go left if i turn right and right if i turn left as if in real life how it works.
public class javafx extends Application {
int ze = 0;
int ye = 0;
int xe = 0;
PerspectiveCamera cam = new PerspectiveCamera();
//the rotation angles//
Rotate rx = new Rotate();
{ rx.setAxis(Rotate.X_AXIS); }
Rotate ry = new Rotate();
{ ry.setAxis(Rotate.Y_AXIS); }
Rotate rz = new Rotate();
{ rz.setAxis(Rotate.Z_AXIS); }
int xt = 0;
int yt = 0;
int one;
int two;
boolean flip = false;
public static void addRotate(Node node, Point3D rotationAxis,
double angle) {
ObservableList<Transform> transforms = node.getTransforms();
try {
for (Transform t : transforms) {
rotationAxis = t.inverseDeltaTransform(rotationAxis);
}
} catch (NonInvertibleTransformException ex) {
throw new IllegalStateException(ex);
}
transforms.add(new Rotate(angle, rotationAxis));
}
public void start(Stage stage) {
Box cube = new Box();
cam.getTransforms().addAll(rx, rz, ry);
cube.setDepth(100.0);
cube.setHeight(100.0);
cube.setWidth(200.0);
cube.setCullFace(CullFace.BACK);
cube.setDrawMode(DrawMode.FILL);
PhongMaterial material = new PhongMaterial();
material.setDiffuseColor(Color.BROWN);
cube.setMaterial(material);
cube.setTranslateX(1500.0);
cube.setTranslateY(500.0);
cube.setTranslateZ(0.0);
cam.setTranslateX(0);
cam.setTranslateY(0);
cam.setTranslateZ(0);
cam.setScaleX(2);
cam.setScaleY(2);
cam.setScaleZ(2);
Group root = new Group(cube, cam);
Dimension dimension =
Toolkit.getDefaultToolkit().getScreenSize();
double screenHeight = dimension.getHeight();
double screenWidth = dimension.getWidth();
Scene scene = new Scene(root, 0,0);
stage.setFullScreen(true);
scene.setCamera(cam);
stage.setTitle("3d space");
stage.setScene(scene);
stage.show();
stage.setOnHiding( event -> { System.exit(0);} );
//to get the input from a mouse vvvv //
ScheduledExecutorService scheduler =
Executors.newScheduledThreadPool(1);
Runnable toRun;
toRun = new Runnable() {
public void run() {
if(!flip){
one = MouseInfo.getPointerInfo().getLocation().x;
two = MouseInfo.getPointerInfo().getLocation().x;
flip = true;
}else{
flip = false;
if(one > MouseInfo.getPointerInfo().getLocation().x){
xt = xt +2;
ry.setAngle(xt);
System.out.println("left");
}else{
if(one < MouseInfo.getPointerInfo().getLocation().x){
System.out.println("right");
xt = xt -2;
ry.setAngle(xt);
}
}
}
}
};
ScheduledFuture<?> handle = scheduler.scheduleAtFixedRate(toRun, 1, 1, TimeUnit.MILLISECONDS);
}
public static void main(String args[]){
launch(args);
}
}

the answer is cam.set.pivot(222) to the area it circles around, also you can rotate everything around the camera

How do I initiate particles to grow from anywhere on stage upon mouseClick?

I want to be able to create new particles from wherever I click on stage. I'm just stuck on adding the mouse portion and I've tried to add/pass parameters and I always get errors while I try setting the parameters. Any suggestions?
This is my current code:
float parSpeed = 1; //speed of particles
int nParticles = 1000; //# of particles
Particle[] particles;
void setup() {
size(700,700); //size of image (8.5 x 11 # 300px) is 3300,2550)
frameRate(60); //framerate of stage
background(0); //color of background
particles = new Particle[nParticles];
//start particle array
for(int i=0; i<nParticles; i++) {
particles[i] = new Particle();
}
}
void draw() {
fill(0,0,0,5); //5 is Alpha
rect(0,0,width,height); //color of rectangle on top?
translate(width/2, height/2); //starting point of particles
//start particle array
for(int i=0; i<nParticles; i++) {
particles[i].update();
particles[i].show();
}
}
//Particle Class
class Particle {
PVector pos; //position
float angle; //angle
float dRange; //diameter range
float dAngle; //beginning angle?
color c; //color
Particle() {
pos = new PVector(0,0);//new position for the particles.
angle = 1; //controls randomization of direction in position when multiplied
dRange = 0.01; // how big of a circle shold the particles rotate on
dAngle = 0.15; // -- maximum angle when starting
c = color(0,0,random(100, 255)); //set color to random blue
}
void update() {
float cor = .25*dRange*atan(angle)/PI;
float randNum = (random(2)-1)*dRange-cor; //Random number from (-dRange, dRange)
dAngle+=randNum; //We don't change the angle directly
//but its differential - source of the smoothness!
angle+=dAngle; //new angle is angle+dAngle -- change angle each frame
pos.x+=parSpeed*cos(angle);//random direction for X axis multiplied by speed
pos.y+=parSpeed*sin(angle);//rabdin durectuib for y axis multiplied by speed
}
void show() {
fill(c); //fill in the random color
noStroke(); //no stroke
ellipse(pos.x,pos.y,10,10); //make the shape
smooth(); //smooth out the animation
}
}
void keyPressed() {
print("pressed " + int(key) + " " + keyCode);
if (key == 's' || key == 'S'){
saveFrame("image-##.png");
}
}
void mouseReleased() {
print("mouse has been clicked!");
}

Overwrite the mouseReleased() method:
In there you'll need to:
capture the position of the mouse
create the new particle
update the position of the newly created particle.
add it to the array (the particle system)
This may look simple but you'll have to keep in mind an array cannot change sizes. I would advise that you create a ParticleSystem class that takes care of adding and removing particles from the system.
Edit: You might want to consider using an ArrayList instead of an Array of Particles. Have a look at this
In pseudo code this would look like this:
void mouseReleased() {
int particleX = mouseX;
int particleY = mouseY;
Particle P = new Particle();
P.setPos ( new PVector ( particleX, particleY ) ); // this needs to be implemented
ParticleSystem.add ( P ); // this needs to be implemented
}
I hope this will be a good start.
AH

Collision Detection in java?

I am currently making a collision detection in java, I have a two objects coming from a different class which is a bird and a worm. The worm moves in a random direction while the bird move only when I want him to move because I'm the one who's controlling him by using key arrow buttons. Now, if the two objects collide I want only the worm to be vanish and the bird will remain there. Could somebody help me how to do that? Thanks in advance for sharing your ideas.
This codes is coming from my World class.
public void render()
{
setupStrategy();
x0_pixel = 0;
y0_pixel = 0;
x1_pixel = getWidth();
y1_pixel = getHeight();
x1_world = x1_pixel / meter;
y1_world = y1_pixel / meter;
Graphics2D g2d = (Graphics2D) strategy.getDrawGraphics();
g2d.setBackground(Color.lightGray);
g2d.clearRect(0, 0, x1_pixel, y1_pixel);
g2d.setColor(Color.BLACK);
for (double x = 0; x < x1_world; x++)
{
for (double y = 0; y < y1_world; y++)
{
int xx = convertToPixelX(x);
int yy = convertToPixelY(y);
g2d.drawOval(xx, yy, 2, 2);
}
}
for (Worm worm : worms)
{
for (Sprite bird : birds)
{
if (!collides())
{
bird.render(g2d);
worm.render(g2d);
}
}
}
g2d.dispose();
strategy.show();
Toolkit.getDefaultToolkit().
sync();
}
public void start() throws InterruptedException
{
long prevLoopStart = System.nanoTime();
Avg avg = new Avg();
while (true)
{
final long loopStart = System.nanoTime();
final long dt = loopStart - prevLoopStart;
for (Worm worm : worms)
{
worm.move(dt);
}
for (Bird bird : birds)
{
bird.move(dt);
}
render();
Collides();
frame.onFpsUpdated(1.0 / dt * SECOND, avg.add(loopStart));
final long elapsed_ns = System.nanoTime() - loopStart;
long expected_elapsed_ms = 1000 / 60;
long elapsed_ms = (long) (elapsed_ns / (1000.0 * 1000.0));
long sleep_ms = expected_elapsed_ms - elapsed_ms;
if (sleep_ms > 0)
{
Thread.sleep(sleep_ms /* ms */);
}
prevLoopStart = loopStart;
}
}
private boolean Collides()
{
ArrayList<Sprite> toDisappear = new ArrayList<Sprite>();
for (int i = 0; i < worms.size(); i++)
{
Sprite r1 = worms.get(i);
for (int j = i + 1; j < birds.size(); j++)
{
Sprite r2 = birds.get(j);
if (r1 == r2)
{
continue;
}
Rectangle me = r1.getBounds();
Rectangle him = r2.getBounds();
if (me.intersects(him))
{
collision = true;
toDisappear.add(r1);
toDisappear.add(r2);
}
toDisappear.remove(r1);
todisappear.remove(r2);
}
}
return collision;
}

This book provides a lot of INFO regarding "Collision Detection" ( Read Chapter 2 and 3 ) and other topics related to Game programming in Java.
http://fivedots.coe.psu.ac.th/~ad/jg/

There are open source frameworks that already do that. If you are doing it for learning purposes then just check their code.
One is Genuts and it SpriteCollisionManager.
The resource that #Upul mention is also of great value.
Regards.

you have a nested for-loop which checks for collisions between birds and worms. if there is a collision neither of them are drawn. Your code also incorrectly draws birds and worms multiple times.
for (Worm worm : worms){
for (Sprite bird : birds){
if (!collides()){
bird.render(g2d);
worm.render(g2d);
}
}
}
The correct way to go is to remove the worm from the list when there is a collision:
for (Iterator<Worm> wit = worms.getIterator(); wit.hasNext(); ){
Worm worm = wit.next();
for (Sprite bird : birds){
if (worm.collidesWith(bird)){
wit.remove();
break;
}
}
}
for (Worm worm : worms){
worm.render(g2d);
}
for (Sprite bird : birds){
bird.render(g2d);
}
Now I assume that you will implement a boolean collidesWith(Bird other) method in class worm so that you can compare that specific worm with the bird you got as an argument. Alternatively make a function as such:
boolean collidesWormBird(Sprite worm, Sprite bird){
Rectangle me = worm.getBounds();
Rectangle him = bird.getBounds();
return me.intersects(him);
}

How to programatically pan a VisualizationViewer with Jung (the java library)?

After a lot a investigations, I don't achieve to find a convenient answer to the following question: how to programatically pan a VisualizationViewer with Jung?
I have a GUI with the list of the vertices of my graph, and I want that a double click on one item of the list (i.e. a node description) perform a "centering action" of my VisualizationViewer for the clicked node.
How to code such a behavior? it seems simple but I found no convenient answer.
If someone could help, thanks!
njames

Here is how to popup a menu after right-clicking on a node in JUNG2 and later choose to center to this node:
graphMouse.add(new MyPopupGraphMousePlugin());
/**
* a GraphMousePlugin that offers popup
* menu support
*/
protected class MyPopupGraphMousePlugin extends AbstractPopupGraphMousePlugin
implements MouseListener {
public MyPopupGraphMousePlugin() {
this(MouseEvent.BUTTON3_MASK);
}
public MyPopupGraphMousePlugin(int modifiers) {
super(modifiers);
}
/**
* If this event is over a node, pop up a menu to
* allow the user to center view to the node
*
* #param e
*/
protected void handlePopup(MouseEvent e) {
final VisualizationViewer<Node,Link> vv =
(VisualizationViewer<Node,Link>)e.getSource();
final Point2D p = e.getPoint();
GraphElementAccessor<Node,Link> pickSupport = vv.getPickSupport();
if(pickSupport != null) {
final Node station = pickSupport.getVertex(vv.getGraphLayout(), p.getX(), p.getY());
if(station != null) {
JPopupMenu popup = new JPopupMenu();
String center = "Center to Node";
popup.add(new AbstractAction("<html><center>" + center) {
public void actionPerformed(ActionEvent e) {
MutableTransformer view = vv.getRenderContext().getMultiLayerTransformer().getTransformer(Layer.VIEW);
MutableTransformer layout = vv.getRenderContext().getMultiLayerTransformer().getTransformer(Layer.LAYOUT);
Point2D ctr = vv.getCenter();
Point2D pnt = view.inverseTransform(ctr);
double scale = vv.getRenderContext().getMultiLayerTransformer().getTransformer(Layer.VIEW).getScale();
double deltaX = (ctr.getX() - p.getX())*1/scale;
double deltaY = (ctr.getY() - p.getY())*1/scale;
Point2D delta = new Point2D.Double(deltaX, deltaY);
layout.translate(deltaX, deltaY);
}
});
popup.show(vv, e.getX(), e.getY());
} else {
}
}
}
}
Edited: Finally! the correct node-to-center-view with scale factor calculation...

public void centerViewsOnVertex(SynsetVertex v) {
//the following location have sense in the space of the layout
Point2D v_location = layout.transform(v);
Point2D vv1_center_location = vv1.getRenderContext()
.getMultiLayerTransformer()
.inverseTransform(Layer.LAYOUT, vv1.getCenter());
double scale = vv1.getRenderContext().getMultiLayerTransformer().getTransformer(Layer.VIEW).getScale();
vv1.getRenderContext().getMultiLayerTransformer().getTransformer(Layer.LAYOUT).translate(
-(v_location.getX() - vv1_center_location.getX()) * 1
/ scale,
-(v_location.getY() - vv1_center_location.getY()) * 1
/ scale);
refreshViews();
}

Since I was just looking for an answer to this and the above worked terribly; here's a code snippet I found in AnimatedPickingGraphMousePlugin that will recenter:
Layout<V,E> layout = vv.getGraphLayout();
Point2D q = layout.transform(vertex);
Point2D lvc =
vv.getRenderContext().getMultiLayerTransformer().inverseTransform(vv.getCenter());
final double dx = (lvc.getX() - q.getX());
final double dy = (lvc.getY() - q.getY());
vv.getRenderContext().getMultiLayerTransformer().getTransformer(Layer.LAYOUT).translate(dx, dy);

Actually, I have found by the next a solution:
//the following location have sense in the space of the layout
Point2D v_location = layout.transform(v);
Point2D vv1_center_location = vv1.getRenderContext()
.getMultiLayerTransformer()
.inverseTransform(Layer.LAYOUT, vv1.getCenter());
vv1.getRenderContext()
.getMultiLayerTransformer()
.getTransformer(Layer.LAYOUT)
.translate(-(v_location.getX() - vv1_center_location.getX()),
-(v_location.getY() - vv1_center_location.getY()));

The center on node function is already implemented in the AnimatedPickingGraphMousePlugin
http://sourceforge.net/p/jung/discussion/252062/thread/18b4b941
In picking mode ctrl+click on a vertex to center on it.
#SuppressWarnings("unchecked")
public void mouseReleased(MouseEvent e) {
if (e.getModifiers() == modifiers) {
final VisualizationViewer<V,E> vv = (VisualizationViewer<V,E>) e.getSource();
if (vertex != null) {
Layout<V,E> layout = vv.getGraphLayout();
Point2D q = layout.transform(vertex);
Point2D lvc = vv.getRenderContext().getMultiLayerTransformer().inverseTransform(vv.getCenter());
final double dx = (lvc.getX() - q.getX()) / 10;
final double dy = (lvc.getY() - q.getY()) / 10;
Runnable animator = new Runnable() {
public void run() {
for (int i = 0; i < 10; i++) {
vv.getRenderContext().getMultiLayerTransformer().getTransformer(Layer.LAYOUT).translate(dx, dy);
try {
Thread.sleep(100);
} catch (InterruptedException ex) {
}
}
}
};
Thread thread = new Thread(animator);
thread.start();
}
}
}

I was trying to figure out how to center on a given vertex and came across this, but unfortunately it was not particularly helpful and I spent a fair amount of time figuring out how to do it. So, sharing my experience here in case it may be helpful for others.
The application I'm writing has a VisualizationViewer, that is loaded inside a GraphZoomScrollPane. I have a GraphMouseListener that I've added to the VisualizationViewer, which allows a user to right click on a vertex in the viewable area of the scroll pane and in the popup menu they can choose to center on the vertex.
The top voted answer on this thread references usage of a MutableTransformer from the LAYOUT layer and it uses the translate method of that transformer to do the centering action. Unfortunately, if you are using a zoom/scroll then you don't really know the size and positioning of the layout layer in relation to the view layer without doing a bunch of extra math.
When using zoom/scroll pane, I'd recommend finding the location of the vertex in the viewable area of the graph as represented by the pane, and then adjusting where the view pane is at.
Here is a snippet of the code I worked out:
void center(MouseEvent me, GraphZoomScrollPane gzsp) {
VisualizationViewer<V,E> vv =
(VisualizationViewer<V,E>)me.getSource();
MutableTransformer viewTransformer =
vv.getRenderContext().getMultiLayerTransformer().getTransformer(Layer.VIEW);
double scaleFromViewTransformer = viewTransformer.getScale();
Dimension paneSize = gzsp.getSize();
Point2D positionOfVertexInPane = me.getPoint();
double[] centerOfPane = new double[] {
paneSize.getWidth()/2d,
paneSize.getHeight()/2d
};
double[] amountToMovePane = new double[] {
(centerOfPane[0]-positionOfVertexInPane.getX())/scaleFromViewTransformer,
(centerOfPane[1]-positionOfVertexInPane.getY())/scaleFromViewTransformer
};
viewTransformer.translate(amountToMovePane[0], amountToMovePane[1]);
}