Following Quote from this source:
http://www.cambridgeincolour.com/tutorials/image-projections.htm
Equirectangular image projections map the latitude and longitude
coordinates of a spherical globe directly onto horizontal and vertical
coordinates of a grid, where this grid is roughly twice as wide as it
is tall.
I have a 13312 px width and 6656 pixel height Panorama picture. It's a equirectangular projection of a room and have a 2:1 ratio.
I use following formular to calculate the xPosition:
var xPosition = ( panorama.width / 360 ) * azimuth
Azimuth = Phi = Heading = Angle to the left or right
How do I project this now on a 1366x768px browser screen?
I think my results are wrong, because it's not on the point where it should be.. it could be because the sphere has a distortion on the left and right:
Is there any formular to calculate the position with attention to the distortion and scale it to fit on the browser screen? I looked many (MANY) sources to find a solution for this, but they always just say that equirectangular are just lat and long.. they don't consider the distortion.
Last question: To find a special solution, I tryed to put a plane on the circle and expanded the line which shows the alpha angle. I though with Phytagoras I could find the position.. but this didn't worked either.. maybe I did something wrong? Is this the way even possible or am I doing it wrong?
edit
THIS is what I'm actually looking for: http://othree.github.io/360-panorama/three-2d/
The black grid in the background. What is the name of this? For what do I have to google or look for? When you start the 2D Panorama, if you want to get the coordinations of the top right corner of the window, what do you have to do?
The whole calculation problem was about to create a Google Streetview similiar view from a 2:1 equirectangular image. We already found a solution for this with a great help from Martin Matysiak (https://github.com/marmat | Google).
It's been a while so I can't give a direct answer to what the main solution is, but I can provide a URL to an AddOn Martin wrote for adding the custom Markers that we actually were trying to make.
You can follow https://github.com/marmat/google-maps-api-addons and look for yourself. In the end it helped a lot to solve the main problem and let us continue with our main Framework for Google Business Tours.
If you follow the link in the threejs demo you included, it would take you to the source code.
particularly look at:
https://github.com/mrdoob/three.js/blob/dev/examples/webgl_panorama_equirectangular.html
and
https://github.com/mrdoob/three.js/blob/dev/src/geometries/SphereBufferGeometry.js
not sure if there is distortion though. The distortion comes from the fact that the texture is mapped to the sphere, and the sphere is rendered in 3D (openGL).
Related
I am using three.js and created a basic scene with a few meshses rendered with some basic lighting . My torus geometry was previously pointy with few polygons but I was able to smooth it out by adding more segments but now I would like to straighten the lines out in my cube. They seem jaggy. In the pic I am including my render on the left seems a lot jaggier then the example I am following on the right. I have even set width and height segments to very large values (200). Any Ideas why the aliasing/jagginess?My render on left, example on right
In a webpage created with node/webpack, vega-lite, and vegaEmbed, I have a layer with rect marks with short annotations inside them using text marks. I'd like to clip the text to its surrounding rect but haven't figured out a way to do this and hope someone can point me in the right direction.
I realize text has a limit property in pixel units. If I could determine the pixel units of my rect marks (I don't know how to do this), using limit seems like a reasonable approach.
Also, if I knew the pixel extents of my rectangle, I can then write code to align the text within the rect which would be desirable. Currently I just use the same x as the rect, with a dx offset.
I've read about background for text which is a similar problem, but not the same.
I have an image of size 480x800 pixels and there is a icon on one corner which I need to place. What I want is that to ignore all touches on the transparent areas and detect only the area where the icon is.
I found a solution in SO to this problem but it just tells the code to be used. I need to know exactly where to put that code since I am a beginner and don't know much about cocos2d so I expect a step by step solution.
Cocos2d 2.0 - Ignoring touches to transparent areas of layers/sprites
Do not use glReadPixels because it affected by bugs in android drivers. You can translate CCTouch to CCPoint in image coordinates using convertTouchToNodeSpace, and read image pixel at given point.
Create CCImage from file that contains semi-transparent picture, and read one pixel at tap point; it should be {0,0,0,0} for transparent area.
Don't forget to check that tap is not outside picture, and create pixel index in CCImage::getData() array with formulae unsigned index = x * imageWidth + y.
I am trying to rotate a circle image using jquery rotate plugin. But, I observe that the circle image appears to be oscillating about it's center while rotating. Something like a linear movement.
In IE7, it's pretty much observable. In Chrome/Firefox, if you zoom a bit, this wobbling issue can be observed.
I tried to rotate the image using Raphael library too. But still, the circle was wobbling.
Then I rotated the image with a c# program to see if the problem is with the browsers' graphic engine. But even here this issue is observed. I even took various images to see if the circle was imperfect. But, all the images show this behavior.
Can this be happening because of anti aliasing? Or, is it just an optical illusion?
If not, then what else can be the reason?
I am using jqueryrotate plugin rotate method to do this. $('#circleimg').rotate(Position);
Check this jsFiddle. Btw, this is the kind of source code you can include yourself in your question to make answering it easier for others.
http://jsfiddle.net/KyJzQ/
I used this rotate plugin since you did not specify which one you used.
HTML
<img id="circleimg" src="http://i.imgur.com/KgUDClr.png">
<img id="circleimg2" src="http://i.imgur.com/sK4qP6z.png">
JAVASCRIPT
var rot = 0;
window.setInterval(function() {
$('#circleimg').rotate(rot);
$('#circleimg2').rotate(rot);
rot += 1;
},
5
);
I think the circle you've provided is not actually a perfect circle, hence the tiny "oscillations" you see.
I added another circle that I rendered myself. If you zoom in closely enough you will see jagged pixels moving around the edge of my circle, but I don't see any way around this since the jQuery plugin is not actually re-rendering the image. I don't see any of the "oscillating" you mention in my circle.
This issue is more about vector art than it is about programming. Get a nice vector graphics program and render the nicest circle you can. Even the one I made is not really that good. Or use HTML5 and canvas to draw dynamically, rather than moving a static image around.
I got a Windows Forms app making use of Google Earth where users can draw a polygon on the map which is used as a geofence.
What I'd like to do is to be able to zoom to the polygon so that it fits nicely on screen with a click of a button. A sort of zoom to fit function.
Finding the centre of the polygon and setting the Google Earth camera to that lat/long is easy.
What I need is an algorithm that takes a bounding box of lats \ longs, screen height \ width and then determines the altitude to set the camera.
Does anyone have this algorithm or know where one can be found?
Thanks!!
To "zoom to the polygon" or any other KmlObject it is probably easiest to use the earth-api-utility-library.
See the methods createBoundsView, setToBoundsView and especially flyToObject
For example, where ge is the GEPlugin object and polygon is a KmlPolygon.
var gex = new GEarthExtensions(ge);
gex.util.flyToObject(polygon, { boundsFallback: true, aspectRatio: 1 } );