What is the advantage of using Bufferedimage for images?

Question

I want to know the difference in using BufferedImage and the method graphics to draw the images to the ImageIcon.

Here is an example:

@Override
    public void paint(Graphics g) {

        g.drawImage(bfImage, 0, 0, this);
        g.drawImage(bfImage2, 10, 350, this);
    }
    public void desenharImagem() {

        try{

            bfImage = ImageIO.read(getClass().getResource("Imagens/cenario.jpg"));
            bfImage2 = ImageIO.read(getClass().getResource("Imagens/0.png"));

        }catch(IOException e) {
            e.printStackTrace();
        }
    }

Answer 1

The interface Image is the one that models the behavior of objects that represent images in Java.

The class BufferedImage is an implementation of Image which corresponds to images represented by a sequence of pixels stored entirely in memory.

The Graphics (as well as his subclass Graphics2D) nay represents an image, and yes is an object that makes the drawing in images.

A certain analogy that would give to do (although very imperfect) is that "the BufferedImage is a role, while the Graphics is a pen".

As you showed yourself, the method to draw an image in a Graphics is as follows:

BufferedImage bi = ...;
g.drawImage(bi, px, py, this);

There is also the method to obtain a Graphics from a BufferedImage, which is the method createGraphics():

Graphics2D g = bi.createGraphics();

Using a BufferedImage, you can even access the pixels individually with the methods getRGB(int, int) and setRGB(int, int, int), but this is usually something very low level to make things like drawings of complex geometric shapes, textures, rendering texts with different colors and fonts in the image, etc. To do this from there manipulating the pixels directly is a hard and difficult work, and that’s where the classes Graphics and Graphics2D enter. They are abstractions that allow you to perform these operations.

The AWT, whenever it calls the method paint(Graphics), it will pass an instance of Graphics2D. Then you can cast it safely. The class Graphics2D offers a much richer set of methods than that offered directly by Graphics.

The interface Icon is intended to be a figure used to decorate Swing components. For example, MetalCheckBoxIcon is the implementation that represents the checkbox square. Already the class ImageIcon is the implementation of Icon which is intended to represent an icon that matches some arbitrary image. Use ImageIcon when only the BufferedImage would already be necessary is conceptually wrong, because the purpose of the class ImageIcon is more to be an adapter between the interfaces Image and Icon.

Ah, and don’t overdo the method paint(Graphics). You should write the paintComponent(Graphics). See more about this in that other answer of mine.