929
iOS中的图像处理（三）——混合运算
有时候，单独对一张图像进行处理是很难或者根本达不到我们想要的效果的。一个好的滤镜效果的诞生，往往要经过很多复杂步骤、细致微调、图片应用效果观察以及很多图层叠加。
我在JSWidget上发现了一些常用混合算法，对应着一些常用混合模式，通过这些blend modes，我们可以指定两张图像如何混合。
不过在此之前，我们需要纯颜色图像和渐变图像来做辅助：
+ (UIImage *)imageWithColor:(UIColor *)color size:(CGSize)size
{
// https://stackoverflow.com/questions/1213790/how-to-get-a-color-image-in-iphone-sdk
//Create a context of the appropriate size
UIGraphicsBeginImageContext(size);
CGContextRef currentContext = UIGraphicsGetCurrentContext();
//Build a rect of appropriate size at origin 0,0
CGRect fillRect = CGRectMake(0, 0, size.width, size.height);
//Set the fill color
CGContextSetFillColorWithColor(currentContext, color.CGColor);
//Fill the color
CGContextFillRect(currentContext, fillRect);
//Snap the picture and close the context
UIImage *colorImage = UIGraphicsGetImageFromCurrentImageContext();
UIGraphicsEndImageContext();
return colorImage;
}
+ (UIImage *)imageWithGradient:(UIImage *)image startColor:(UIColor *)startColor endColor:(UIColor *)endColor
{
UIGraphicsBeginImageContextWithOptions(image.size, NO, image.scale);
CGContextRef context = UIGraphicsGetCurrentContext();
CGContextTranslateCTM(context, 0, image.size.height);
CGContextScaleCTM(context, 1.0, -1.0);
CGContextSetBlendMode(context, kCGBlendModeNormal);
CGRect rect = CGRectMake(0, 0, image.size.width, image.size.height);
CGContextDrawImage(context, rect, image.CGImage);
// Create gradient
NSArray *colors = [NSArray arrayWithObjects:(id)endColor.CGColor, (id)startColor.CGColor, nil];
CGColorSpaceRef space = CGColorSpaceCreateDeviceRGB();
CGGradientRef gradient = CGGradientCreateWithColors(space, (CFArrayRef)colors, NULL);
// Apply gradient
CGContextClipToMask(context, rect, image.CGImage);
CGContextDrawLinearGradient(context, gradient, CGPointMake(0,0), CGPointMake(0, image.size.height), 0);
UIImage *gradientImage = UIGraphicsGetImageFromCurrentImageContext();
UIGraphicsEndImageContext();
CGGradientRelease(gradient);
CGColorSpaceRelease(space);
return gradientImage;
}
而且在第一篇文章中提到的透明度滤镜（作用域像素的alpha值上）是没效果的，可以通过Quartz 2D来实现：
- (UIImage *)setAlpha:(CGFloat)alpha
{
// https://stackoverflow.com/questions/5084845/how-to-set-the-opacity-alpha-of-a-uiimage
UIGraphicsBeginImageContextWithOptions(self.size, NO, 0.0f);
CGContextRef ctx = UIGraphicsGetCurrentContext();
CGRect area = CGRectMake(0, 0, self.size.width, self.size.height);
CGContextScaleCTM(ctx, 1, -1);
CGContextTranslateCTM(ctx, 0, -area.size.height);
CGContextSetBlendMode(ctx, kCGBlendModeMultiply);
CGContextSetAlpha(ctx, alpha);
CGContextDrawImage(ctx, area, self.CGImage);
UIImage *newImage = UIGraphicsGetImageFromCurrentImageContext();
UIGraphicsEndImageContext();
return newImage;
}
在此基础上，通过下面四行代码，可以分别得到四种不同效果：
return [[UIImage imageWithColor:[UIColor purpleColor] size:originImage.size] changeOpacityByFactor:0.5];;
return [UIImage imageWithGradient:originImage startColor:[UIColor whiteColor] endColor:[UIColor yellowColor]];
return [[originImage tintWithMaxRGBA:(RGBA){190, 190, 230} minRGBA:(RGBA){50, 35, 10}] overlayWithImage:[[UIImage imageWithColor:[UIColor purpleColor] size:originImage.size] changeOpacityByFactor:0.3]];
return [originImage softlightWithImage:[[UIImage imageWithColor:[UIColor yellowColor] size:originImage.size] changeOpacityByFactor:0.8]];
其中，overlay算法如下：
double calcOverlay(float b, float t)
{
return (b > 128.0f) ? 255.0f - 2.0f * (255.0f - t) * (255.0f - b) / 255.0f: (b * t * 2.0f) / 255.0f;
}
void filterOverlay(UInt8 *pixelBuf, UInt8 *pixedBlendBuf, UInt32 offset, void *context)
{
int r = offset;
int g = offset+1;
int b = offset+2;
int red = pixelBuf[r];
int green = pixelBuf[g];
int blue = pixelBuf[b];
int blendRed = pixedBlendBuf[r];
int blendGreen = pixedBlendBuf[g];
int blendBlue = pixedBlendBuf[b];
pixelBuf[r] = SAFECOLOR(calcOverlay(red, blendRed));
pixelBuf[g] = SAFECOLOR(calcOverlay(green, blendGreen));
pixelBuf[b] = SAFECOLOR(calcOverlay(blue, blendBlue));
}
最后更新：2017-04-02 16:48:10