这可以通过RadialGradientPaint 和适当的AlphaComposite 来解决。
以下是MCVE,它显示了如何做到这一点。它使用与user1803551 used in his answer 相同的图像,因此屏幕截图看起来(几乎)相同。但是这个添加了一个MouseMotionListener,它允许您通过将当前鼠标位置传递给updateGradientAt方法来移动孔,在该方法中实际创建所需的图像:
- 它首先用原始图像填充图像
- 然后它创建一个
RadialGradientPaint,它的中心颜色完全不透明,边框颜色完全透明(!)。这可能看起来违反直觉,但其目的是从现有图像中“切出”洞,这是通过下一步完成的:
-
AlphaComposite.DstOut 分配给Graphics2D。这会导致 alpha 值的“反转”,如公式中所示
Ar = Ad*(1-As)
Cr = Cd*(1-As)
其中r 代表“结果”,s 代表“来源”,d 代表“目标”
结果是图像在所需位置具有径向渐变透明度,在中心完全透明,在边界完全不透明 (!)。然后使用Paint 和Composite 的组合来填充具有孔的大小和坐标的椭圆。 (也可以进行fillRect 调用,填充整个图像 - 它不会改变结果)。
import java.awt.AlphaComposite;
import java.awt.Color;
import java.awt.Graphics;
import java.awt.Graphics2D;
import java.awt.Point;
import java.awt.RadialGradientPaint;
import java.awt.event.MouseAdapter;
import java.awt.event.MouseEvent;
import java.awt.image.BufferedImage;
import java.io.File;
import java.io.IOException;
import javax.imageio.ImageIO;
import javax.swing.JFrame;
import javax.swing.JPanel;
import javax.swing.SwingUtilities;
public class TransparentGradientInImage
{
public static void main(String[] args)
{
SwingUtilities.invokeLater(new Runnable()
{
@Override
public void run()
{
createAndShowGUI();
}
});
}
private static void createAndShowGUI()
{
JFrame f = new JFrame();
f.setDefaultCloseOperation(JFrame.EXIT_ON_CLOSE);
TransparentGradientInImagePanel p =
new TransparentGradientInImagePanel();
f.getContentPane().add(p);
f.setSize(800, 600);
f.setLocationRelativeTo(null);
f.setVisible(true);
}
}
class TransparentGradientInImagePanel extends JPanel
{
private BufferedImage background;
private BufferedImage originalImage;
private BufferedImage imageWithGradient;
TransparentGradientInImagePanel()
{
try
{
background = ImageIO.read(
new File("night-sky-astrophotography-1.jpg"));
originalImage = convertToARGB(ImageIO.read(new File("7bI1Y.jpg")));
imageWithGradient = convertToARGB(originalImage);
}
catch (IOException e)
{
e.printStackTrace();
}
addMouseMotionListener(new MouseAdapter()
{
@Override
public void mouseMoved(MouseEvent e)
{
updateGradientAt(e.getPoint());
}
});
}
private void updateGradientAt(Point point)
{
Graphics2D g = imageWithGradient.createGraphics();
g.drawImage(originalImage, 0, 0, null);
int radius = 100;
float fractions[] = { 0.0f, 1.0f };
Color colors[] = { new Color(0,0,0,255), new Color(0,0,0,0) };
RadialGradientPaint paint =
new RadialGradientPaint(point, radius, fractions, colors);
g.setPaint(paint);
g.setComposite(AlphaComposite.DstOut);
g.fillOval(point.x - radius, point.y - radius, radius * 2, radius * 2);
g.dispose();
repaint();
}
private static BufferedImage convertToARGB(BufferedImage image)
{
BufferedImage newImage =
new BufferedImage(image.getWidth(), image.getHeight(),
BufferedImage.TYPE_INT_ARGB);
Graphics2D g = newImage.createGraphics();
g.drawImage(image, 0, 0, null);
g.dispose();
return newImage;
}
@Override
protected void paintComponent(Graphics g)
{
super.paintComponent(g);
g.drawImage(background, 0, 0, null);
g.drawImage(imageWithGradient, 0, 0, null);
}
}
您可以使用RadialGradientPaint 的fractions 和colors 来实现不同的效果。例如,这些值...
float fractions[] = { 0.0f, 0.1f, 1.0f };
Color colors[] = {
new Color(0,0,0,255),
new Color(0,0,0,255),
new Color(0,0,0,0)
};
造成一个小的透明孔,带有一个大而柔软的“电晕”:
而这些值
float fractions[] = { 0.0f, 0.9f, 1.0f };
Color colors[] = {
new Color(0,0,0,255),
new Color(0,0,0,255),
new Color(0,0,0,0)
};
产生一个大而清晰的中心,带有一个小的“电晕”:
RadialGradientPaint JavaDocs 有一些示例可能有助于找到所需的值。
我发布(类似)答案的一些相关问题:
编辑针对在 cmets 中提出的有关性能的问题
Paint/Composite 方法的性能与getRGB/setRGB 方法相比如何的问题确实很有趣。根据我之前的经验,我的直觉是第一个比第二个快得多,因为一般来说,getRGB/setRGB 往往很慢,并且内置机制是高度优化的(并且,在某些情况下,甚至可能是硬件加速的)。
事实上,Paint/Composite 方法比 getRGB/setRGB 方法快,但不如我预期的快。以下当然不是一个真正深刻的“基准”(我没有为此使用 Caliper 或 JMH),但应该对实际性能给出一个很好的估计:
// NOTE: This is not really a sophisticated "Benchmark",
// but gives a rough estimate about the performance
import java.awt.AlphaComposite;
import java.awt.Color;
import java.awt.Graphics2D;
import java.awt.Point;
import java.awt.RadialGradientPaint;
import java.awt.image.BufferedImage;
public class TransparentGradientInImagePerformance
{
public static void main(String[] args)
{
int w = 1000;
int h = 1000;
BufferedImage image0 = new BufferedImage(w, h,
BufferedImage.TYPE_INT_ARGB);
BufferedImage image1 = new BufferedImage(w, h,
BufferedImage.TYPE_INT_ARGB);
long before = 0;
long after = 0;
int runs = 100;
for (int radius = 100; radius <=400; radius += 10)
{
before = System.nanoTime();
for (int i=0; i<runs; i++)
{
transparitize(image0, w/2, h/2, radius);
}
after = System.nanoTime();
System.out.println(
"Radius "+radius+" with getRGB/setRGB: "+(after-before)/1e6);
before = System.nanoTime();
for (int i=0; i<runs; i++)
{
updateGradientAt(image0, image1, new Point(w/2, h/2), radius);
}
after = System.nanoTime();
System.out.println(
"Radius "+radius+" with paint "+(after-before)/1e6);
}
}
private static void transparitize(
BufferedImage imgA, int centerX, int centerY, int r)
{
for (int x = centerX - r; x < centerX + r; x++)
{
for (int y = centerY - r; y < centerY + r; y++)
{
double distance = Math.sqrt(
Math.pow(Math.abs(centerX - x), 2) +
Math.pow(Math.abs(centerY - y), 2));
if (distance > r)
continue;
int argb = imgA.getRGB(x, y);
int a = (argb >> 24) & 255;
double factor = distance / r;
argb = (argb - (a << 24) + ((int) (a * factor) << 24));
imgA.setRGB(x, y, argb);
}
}
}
private static void updateGradientAt(BufferedImage originalImage,
BufferedImage imageWithGradient, Point point, int radius)
{
Graphics2D g = imageWithGradient.createGraphics();
g.drawImage(originalImage, 0, 0, null);
float fractions[] = { 0.0f, 1.0f };
Color colors[] = { new Color(0, 0, 0, 255), new Color(0, 0, 0, 0) };
RadialGradientPaint paint = new RadialGradientPaint(point, radius,
fractions, colors);
g.setPaint(paint);
g.setComposite(AlphaComposite.DstOut);
g.fillOval(point.x - radius, point.y - radius, radius * 2, radius * 2);
g.dispose();
}
}
我的电脑上的时间是这样的
...
Radius 390 with getRGB/setRGB: 1518.224404
Radius 390 with paint 764.11017
Radius 400 with getRGB/setRGB: 1612.854049
Radius 400 with paint 794.695199
表明Paint/Composite 方法的速度大约是getRGB/setRGB 方法的两倍。除了性能之外,Paint/Composite 还有一些其他优势,主要是上面描述的RadialGradientPaint 的可能参数化,这就是我更喜欢这个解决方案的原因。