How to scale an image using openGL (JGL)?

I want to scale an image using openGL, c开发者_如何转开发an anyone provide me with such code about how to do this ?

PS, I am using JGL as an openGL library for Java.

---

I will be brief on this, as you can find tutorials for pretty much every part of the solution.

You need to load your image from the disk, and create a texture.

You need to create a framebuffer object (FBO) with the desired target dimensions (in your case, double width, double height). Make the FBO active for rendering.

Render a fullscreen quad with your texture applied.

Read the result using glReadPixels().

And that's it ... the texturemapping hardware will take care of rescaling it for you. But it will likely be slower than if done on CPU, especially for "scaling 2x".

EDIT: as requested by OP, it is necessary to include source code.

So ... for loading an image in Java, you would go for this:

BufferedImage img;
try {
    img = ImageIO.read(new File("myLargeImage.jpg"));
} catch (IOException e) { /* ... */ }
int w = img.getWidth(), h = img.getHeight();

For JGL, one might want to convert an image to an array:

byte [][][] imageData = new byte[img.getWidth()][img.getHeight()][3]; // alloc buffer for RGB data
for(int y = 0; y < h; ++ y) {
    for(int x = 0; x < w; ++ x) {
        int RGBA = img.getRGB(x, y);
        imageData[x][y][0] = RGBA & 0xff;
        imageData[x][y][1] = (RGBA >> 8) & 0xff;
        imageData[x][y][2] = (RGBA >> 16) & 0xff;
    }
}

Note that there could be alpha channel as well (for transparency) and that this will likely be quite slow. One could also use:

int[] rgbaData = img.GetRGB(0, 0, w, h, new int[w * h], 0, w);

But that doesn't return the data in the correct format expected by JGL. Tough luck. Then you need to fill a texture:

int[] texId = {0};
gl.glGenTextures(1, texId); // generate texture id for your texture (can skip this line)
gl.glEnable(GL.GL_TEXTURE_2D);
gl.glBindTexture(GL.GL_TEXTURE_2D, texId[0]); // bind the texture
gl.glPixelStorei(GL.GL_UNPACK_ALIGNMENT, 1); // set alignment of data in memory (a good thing to do before glTexImage)
gl.glTexParameteri(GL.GL_TEXTURE_2D, GL.GL_TEXTURE_WRAP_S, GL.GL_CLAMP);
gl.glTexParameteri(GL.GL_TEXTURE_2D, GL.GL_TEXTURE_WRAP_T, GL.GL_CLAMP); // set clamp (GL_CLAMP_TO_EDGE would be better)
gl.glTexParameteri(GL.GL_TEXTURE_2D, GL.GL_TEXTURE_MIN_FILTER, GL.GL_LINEAR);
gl.glTexParameteri(GL.GL_TEXTURE_2D, GL.GL_TEXTURE_MAG_FILTER, GL.GL_LINEAR); // set linear filtering (so you can scale your image)
gl.glTexImage2D(GL.GL_TEXTURE_2D, 0, GL.GL_RGB, w, h, 0, GL.GL_RGB, GL.GL_UNSIGNED_BYTE, imageData); // upload image data to the texture

Once you have a texture, you can draw stuff. Let's resample your image:

int newW = ..., newH = ...; // fill-in your values
gl.glViewport(0, 0, newW, newH); // set viewport

gl.glMatrixMode(GL.GL_MODELVIEW);
gl.glLoadIdentity();
gl.glMatrixMode(GL.GL_PROJECTION);
gl.glLoadIdentity();
gl.glColor3f(1.0f, 1.0f, 1.0f); // set "white" color
gl.glDisable(GL.GL_CULL_FACE); // disable backface culling
gl.glDisable(GL.GL_LIGHTING); // disable lighting
gl.glDisable(GL.GL_DEPTH_TEST); // disable depth test
// setup OpenGL so that it renders texture colors without modification (some of that is not required by default)

gl.glBegin(GL_QUADS);
gl.glTexCoord2f(0.0f, 1.0f);
gl.glVertex2f(-1.0f, -1.0f);
gl.glTexCoord2f(1.0f, 1.0f);
gl.glVertex2f(+1.0f, -1.0f);
gl.glTexCoord2f(1.0f, 0.0f);
gl.glVertex2f(+1.0f, +1.0f);
gl.glTexCoord2f(0.0f, 0.0f);
gl.glVertex2f(-1.0f, +1.0f);
gl.glEnd();
// draw a fullscreen quad with your texture on it (scaling is performed here)

Now that the scaled image is rendered, all that needs to be done, is to download it.

byte[][][] newImageData = new byte[newW][newH][3];
gl.glPixelStorei(GL.GL_PACK_ALIGNMENT, 1); // set alignment of data in memory (this time pack alignment; a good thing to do before glReadPixels)
gl.glReadPixels(0, 0, newW, newH, GL.GL_RGB, GL.GL_UNSIGNED_BYTE, newImageData);

And then the data can be converted to BufferedImage in similar way the input image img was converted to imageData.

Note that I used variable gl, that is an instance of the GL class. Put the code into some JGL example and it should work.

One word of caution, JGL doesn't seem to support framebuffer objects, so that you are limited by output image size by your OpenGL window size (attempting to create larger images will result in black borders). It can be solved by using multipass rendering (render your image tile by tile and assemble the full image in memory).