Opencv Java adjusting rectangle dimensions - java

I am working on licence plate detection in OpenCv, currently I can detect a licence plate Sample of detected licence plate
But the issue is the rectangle is too close to the licence plate characters, what I thought was I could just increase the dimensions by a given offset of which I did Sample of increased offset detection
But unfortunately my understanding of the Rectis different from how it works, unlike the circle where you have a single point where it's drawn from, the rectangle uses 2 points, of which after increasing the dimensions, if shifts to the right(At least that is what it seems), I need help on centering the rectangle on its original location after increasing the offset, here is the code am using to increase it's dimensions
rect.height = (int) (rect.height * 1.1);
rect.width = (int) (rect.width * 1.5);
Imgproc.rectangle(originalFrame, rect.br(), rect.tl(), new Scalar(0,0,255), 2);
I don't have enough rep yet so images do not display automatically.

If you increase the right side's x coordinate by length, then you need to make all of the rectangle's coordinates go left length/2. The same applies to height.
rect.xCoord = rect.xCoord - ((rect.width * 1.1) - rect.width) / 2)
This goes before the first line of code you posted above. I'm not sure if this is how you access x coordinates in open cv (since I don't know it), so replace the x coord access with the actual one if this is wrong.

Related

Java Mandelbrot visualization questions on zooming and coloring

I am trying to program a visualisation for the Mandelbrot set in java, and there are a couple of things that I am struggling with to program. I realize that questions around this topic have been asked a lot and there is a lot of documentation online but a lot of things seem very complicated and I am relatively new to programming.
The first issue
The first issue I have is to do with zooming in on the fractal. My goal is to make an "infinite" zoom on the fractal (of course not infinite, as far as a regular computer allows it regarding calculation time and precision). The approach I am currently going for is the following on a timer:
Draw the set using some number of iterations on the range (-2, 2) on the real axis and (2, 2) on the imaginary axis.
Change those ranges to zoom in.
Redraw that section of the set with the number of iterations.
It's the second step that I struggle with. This is my current code:
for (int Py = beginY; Py < endY; Py++) {
for (int Px = beginX; Px < endX; Px++) {
double x0 = map(Px, 0, height,-2, 2);
double y0 = map(Py, 0, width, -2, 2);
Px and Py are the coordinates of the pixels in the image. The image is 1000x1000. The map funtion takes a number, in this case Px or Py, with a range of (0, 1000) and devides it evenly over the range (-2, 2), so it returns the corresponding value in that range.
I think that in order to zoom in, I'll have to change the -2 and 2 values by some way in the timer, but whatever I try, it doesn't seem to work. The zoom always ends up slowing down after a while or it will end up zooming in on a part of the set that is in the set, so not the borders. I tried multiplying them by some scale factor every timer tick, but that doesn't really produce the result I was looking for.
Now I have two questions about this issue.
Is this the right approach to visualizing the set and zooming in(draw, change range, redraw)?
If it is, how do I zoom in properly on an area that is interesting and that will keep zooming in properly even after running for a minute?
The second issue
Of course when visualizing something, you need to get some actual visual thing. In this case I want to color the set in a way similar to what you see here: (https://upload.wikimedia.org/wikipedia/commons/f/fc/Mandel_zoom_08_satellite_antenna.jpg).
My guess is that you have use the amount of iterations a pixel went through to before breaking out of the loop to give it some color value. However, I only really know how to do this with a black and white color scheme. I tried making a color array that holds the same amount of different gray colors as the amount of max iterations, starting from black and ending in white. Here is my code:
Color[] colors = new Color[maxIterations + 2];
for (int i = 0; i < colors.length; i++) {
colors[i] = new Color((int)map(i, 0, maxIterations + 2, 0, 255),
(int)map(i, 0, maxIterations + 2, 0, 255),
(int)map(i, 0, maxIterations + 2, 0, 255));
}
I then just filled in the amount of iterations in the array and assigned that color to the pixel. I have two questions about this:
Will this also work as we zoom into the fractal in the previously described manner?
How can I add my own color scheme in this, like in the picture? I've read some things about "linear interpolation" but I don't really understand what it is and in what way it can help me.
It sounds like you've made a good start.
Re the first issue: I believe there are ways to automatically choose an "interesting" portion of the set to zoom in on, but I don't know what they are. And I'm quite sure it involves more than just applying some linear function to your current bounding rectangle, which is what it sounds like you're doing.
So you could try to find out what these methods are (might get mathematically complicated), but if you're new to programming, you'll probably find it easier to let the user choose where to zoom. This is also more fun in the beginning, since you can run your program repeatedly and explore a new part of the set each time.
A simple way to do this is to let the user draw a rectangle over the image, and use your map function to convert the pixel coordinates of the drawn rectangle to the new real and imaginary coordinates of your zoom area.
You could also combine both approaches: once you've found somewhere you find interesting by manually selecting the zoom area, you can set this as your "final destination", and have the code gradually and smoothly zoom into it, to create a nice movie.
It will always get gradually slower though, as you start using ever more precise coordinates, until you reach the limits of precision with double and it becomes a pixellated mess. From there, if you want to zoom further, you'll have to look into arbitrary-precision arithmetic with BigDecimal - and it will continue to get slower and slower.
Re the second issue: starting off by calculating a value of numIterations / maxIterations (i.e. between 0 and 1) for each pixel is the right idea (I think this is basically what you're doing).
From there, there are all sorts of ways to convert this value to a colour, it's time to get creative!
A simple one is to have an array of a few very different colours. E.g. if you had white (0.0), red (0.25), green (0.5), blue (0.75), black (1.0), then if your calculated number was exactly one of the ones listed, you'd use the corresponding colour. If it's somewhere between, you blend the colours, e.g. for 0.3 you'd take:
((0.5-0.3)*red + (0.3-0.25)*green) / (0.5 - 0.25)
= 0.8*red + 0.2*green
Taking a weighted average of two colours is something I'll leave as an exercise ;)
(hint: take separate averages of the r, g, and b values. Playing with the alpha values could maybe also work).
Another one, if you want to get more mathsy, is to take an equation for a spiral and use it to calculate a point on a plane in HSB colour space (you can keep the brightness at some fixed value, say 1). In fact, any curve in 2D or 3D which you know how to write as an equation of one real variable can be used this way to give you smoothly changing colours, if you interpret the coordinates as points in some colour space.
Hope that's enough to keep you going! Let me know if it's not clear.

Java planetary orbit simulation: centering planets

I've created a simple planetary simulation where a planet orbits a star.
The code for the orbit is this:
a = a + vel * delta;
planetX = Math.cos(a) * orbitRadius + parentStar.getX();
planetY = Math.sin(a) * orbitRadius + parentStar.getY();
Now that works just fine, but my problem is that the orbit is not from the center of the planet around the center of the star.
This is what happens
As you can see, the first red dot on the small circle is the Position of the planet wich orbits around the second small red dot, this is because the circle is drawn from (0,0), so both the planets (0,0) circles around the (0,0) of the star.
I need the the center of the planet to circle the stars center, not their origin point.
Is there a good fix for this?
Your calculation of the orbit is fine. The only problem seems to be that you treat "position" differently when calculating orbits and when drawing the planets: When you draw them, you treat x and y as one of the corner points, but when you calculate the oribit, you treat them as the centre of the body. The simplest way would be to change the visualisation, not the calculation.
Since you did not post the code you use to draw the shapes, I can only guess, but I assume it looks somewhat like this (obviously Pseudocode):
for (Planet p : starsAndPlanets) {
drawCircle(p.x, p.y, p.radius * 2, p.radius * 2);
}
Change this to something like this:
for (Planet p : starsAndPlanets) {
drawCircle(p.x - p.radius, p.y - p.radius, p.radius * 2, p.radius * 2);
}
This way, x and y are the position of the centre of the planet, and with p.x - p.radius and p.y - p.radius you get the corner point. Of course, you could in a similar way change all your orbital mechanic formulas to calculate the centre from the corner point, but IMHO it is much simpler and more natural to treat x and y as the centre.
For now the most suitable way I can think of is getting the star's world coordnates and passing them every frame to the child's coordinates. As you do so, the child would have the same coordinates everyframe.
The next part is translating it and rotating it around the Star - the way you can achieve that is by setting the planet's position to be transposed by the Star's position with a sin(x)*cos(x).
Let me show you an example:
planet[0] = star[0] + sin(angle)*scale
planet[1] = star[1] + cos(angle)*scale
Where the angle would change incrementally and the scale will just shift the child object further from its parent, keeping it a constant (or modifying it if you wish) thus increasing the radius from its 'new' center.
I know some people may mention matrices or other types of transformations, but for this situation I think the above solution would be most relevant and cleanest in my opinionp
The way it works is you take the parent's 'WORLD coordinates' and set them to be the child's. By modifying the Scale value you increase the distance of the object from the center (so they won't overlap) and you multiply this with the sin and cos of the angle you specified to make it rotate.
P.S. Keep in mind that if you're dealing an FPS-dependant engine to render, the more FPS the faster the simulation will be, and vice-versa, because if you render at 1000 fps, this means you execute your code 1000 times, compared to 100 for example. Therefore, you will increment the angle 1000 times or 100 respectively. If you have this issue, try setting a constant framerate if you can - it's the simplest workaround for lightweight simulations.
Edit: I forgot to mention that the concept works for all objects in your case. You just have to work our the relationships and use the function for eqch object seperately where each object has a position and angle of orbit (if it orbits around a different object).

Rotating and drawing an image at a point in Java

I tried to post a different topic for this but people didn't really seem to understand what I was trying to do, so, now I've closed that one and opened this one to give more detail and rephrase the question as a whole.
Ok.
So basically, I have an application which draws an Ellipse. Now, I have a certain number of points (that can be random) in which I have to rotate an image and draw at.
Using Maths I know that to get a point on an Ellipse based by using an angle I use the following equation;
final int radiusW = (width / 2);
final int radiusH = (height / 2);
final int angle = 120;
int pointX = (int) (radiusW + (radiusW * Math.cos(Math.toRadians(angle))));
int pointY = (int) (radiusH + (radiusH * Math.sin(Math.toRadians(angle))));
And that works fine, I can locate an absolute point around the perimeter of the Ellipse.
However, now I'm trying to draw an image on this point so that the image is rotated facing the center of ellipse and is centered on the point.
So, to get the image rotated to the center of the point I do the following;
final AffineTransform at = new AffineTransform();
at.rotate(Math.toRadians(angle - 90), image.getWidth() / 2, image.getHeight() / 2);
final AffineTransformOp ato = new AffineTransformOp(at, AffineTransformOp.TYPE_BILINEAR);
And then I get my new rotated BufferedImage using;
BufferedImage rotated = ato.filter(image, null);
However, I can't seem to be able to get the central point of the image anymore.
If the angle was 0 so that the Image was the original direction then I would simply do;
graphics.drawImage(rotated, pointX - rotated.getWidth() / 2, pointY - rotated.getHeight() / 2, this);
However I'm not sure how to find the central point and draw it based upon that on a rotated image.
I know it involves using cos and sin to multiply the original pointX and pointY by the rotation matrix but everytime I try and work out a solution it always draws completely wrong.
Any help would be very much appreciated as I've spent the best part of a day trying to resolve this.
Thank you.
The thing is that if you just use sin and cos to rotate the corner of the image you will end up with the new rotated position of that corner - when actually what you want to find is the new width and height.
The center is width/2, height/2
Use this to calculate your new width and height:
Calculate Bounding box coordinates from a rotated rectangle
I have done this with OpenCV, and there the image have been rotated but the resulted image was having the same width and height as the initial on. The image was cropped if it was getting out of the initial dimentions and there were black pixels if no information (the rotated image has no pixels in that place).
If you think that rotated has more pixels than the initial image you can verify with size() or length(). Or you can play with the diagonals of the initial rectangle (image size): compute the projection of the diagonals and thake the greatest or what you think. But I am sure that it is similar to the OpenCV case.
I don't know if this can help you, but I hope so.
Your question is not entirely clear but it seems to me that you have a misunderstanding of the nature of the angle used to parametrise your ellipse. The angle as you have it is merely used to parametrise the form of an ellipse equation. It is not the same as the polar angle (except at particular angles). That is to say if you evaluate a point on your ellipse using an angle of (pi/4) radians (45 degrees), then measured the angle that the line from the ellipse centre to your point makes with the axis, it will not measure 45 degrees (except for the case where the ellipse is actually a circle).
That is to say that
int pointX = (int) (radiusW + (radiusW * Math.cos(Math.toRadians(angle))));
int pointY = (int) (radiusH + (radiusH * Math.sin(Math.toRadians(angle))));
is just a parametrisation of an ellipse and that angle is not a polar angle and treating this angle as a rotation angle will not give accurate results (except at integer multiples of (pi/2) radians)
It seems to me that you require the polar form of an ellipse relative to its centre in order for your code to make sense in the context of using this angle for rotation.
It is also possible that I have misunderstood your question though, in which case this answer will be downvoted on a grand scale and I will delete it.

Find the size of the center of a doughnut chart - achartengine

I want to have clickable regions inside of my doughnut chart using aChartEngine, but this is not yet and implemented feature. I think I have a way to implement this so I'm planning out the pieces I need. The one part that I'm not sure about is how to get the size of the hole in the center of the chart. Specifically, I need to know the inner radius to ignore clicks in that area. Does someone know how this is decided? I tried looking through some of the source code but I couldn't quite find what I needed. Let me know if I can add any more information.
You can reverse engineer it from https://code.google.com/p/achartengine/source/browse/trunk/achartengine/src/org/achartengine/chart/DoughnutChart.java:
Basically the radius of the inner circle can be calculated as follows:
int shortestSide = Math.min(Math.abs(right - left), Math.abs(bottom - top));
int originalRadius = (int) (0.35 * shortestSide);
int finalRadius = (int) (originalRadius - (0.2 * shortestSide + 1 * numCategories));
Where final radius is the radius of the inner circle. At least that works for me.

scale() and shape() in Processing

UPDATE: I think I figured it out. The scaleMultiplier also applied to the translate that I was trying to do. I had a suspicion this was the case but couldn't figure out exactly how it got affected.
Anyone familiar with a spinoff of Java called Processing? I'm trying to do something simple, scale a shape and place it in the center of the sketch. This is my code in a nutshell:
pushMatrix();
float scaleX, scaleY, scaleMultiplier, resetX, resetY, transX, transY;
scaleX = 500 / (float)clickState.bounds.getWidth();
scaleY = 500 / (float)clickState.bounds.getHeight();
scaleMultiplier = min(scaleX,scaleY);
resetX = -(float)clickState.bounds.getX();
resetY = -(float)clickState.bounds.getY();
transX = resetX + ((800 - ((float)clickState.bounds.getWidth() * scaleMultiplier))/2);
transY = resetY + ((550 - ((float)clickState.bounds.getHeight() * scaleMultiplier))/2);
scale(scaleMultiplier);
shape(clickState.pshape, transX, transY);
popMatrix();
What I'm trying to do is scale a state on a US Map. clickState is the state that the user clicked on. clickState.bounds is a Rectangle that surrounds the shape of the state. getX() and getY() return the x and y coords of the upper left hand corner of said box. I want to scale the state so that it's no bigger than 500x500 pixels. After that, I want to translate it so that it's in the middle of the sketch, which is 800x550.
Here's where I run into trouble: When I do
shape(clickState.pshape, resetX, resetY);
It draws the state in the upper left hand corner of the sketch. That's exactly what it should do. Then from there I want to add the number of pixels that it would take to center the shape, which is what transX and transY are for. However, when I use transX and transY, the shape gets drawn almost completely off the canvas, depending on where the state is located (even if its only being moved 50 pixels to the right from 0,0). It doesn't make any sense to me whatsoever. I've tried all sorts of combinations of translate(), and even skipping scale() altogether and using a width and height in shape(). It's like Processing is trying to frustrate me on purpose. Any help is appreciated.
I'm not familiar with Processing, but matrix operations are not commutative—the order matters. Typically, operations are applied in an apparent last-specified-first-applied order, as shown in this example.
Vince, I think I was trying to do something very similar last week. Sounds like you figured it out, but check out the answers on this board:
http://forum.processing.org/topic/how-do-i-move-svg-child-shapes#25080000000689051

Categories

Resources