Cropping an UIImage

I've got some code that resizes an image so I can get a scaled chunk of the center of the image - I use this to take a UIImage and return a small, square representation of an image, similar to what's seen in the album view of the Photos app. (I know I could use a UIImageView and adjust the crop mode to achieve the same results, but these images are sometimes displayed in UIWebViews).

I've started to notice some crashes in this code and I'm a bit stumped. I've got two different theories and I'm wondering if either is on-base.

Theory 1) I achieve the cropping by drawing into an offscreen image context of my target size. Since I want the center portion of the image, I set the CGRect argument passed to drawInRect to something that's larger than the bounds of my image context. I was hoping this was Kosher, but am I instead attempting to draw over other memory that I shouldn't be touching?

Theory 2) I'm doing all of this in a background thread. I know there are portions of UIKit that are restricted to the main thread. I was assuming / hoping that drawing to an offscreen view wasn't one of these. Am I wrong?

(Oh, how I miss NSImage's drawInRect:fromRect:operation:fraction: method.)

Update 2014-05-28: I wrote this when iOS 3 or so was the hot new thing, I'm certain there are better ways to do this by now, possibly built-in. As many people have mentioned, this method doesn't take rotation into account; read some additional answers and spread some upvote love around to keep the responses to this question helpful for everyone.

Original response:

I'm going to copy/paste my response to the same question elsewhere:

There isn't a simple class method to do this, but there is a function that you can use to get the desired results: CGImageCreateWithImageInRect(CGImageRef, CGRect) will help you out.

Here's a short example using it:

CGImageRef imageRef = CGImageCreateWithImageInRect([largeImage CGImage], cropRect);
// or use the UIImage wherever you like
[UIImageView setImage:[UIImage imageWithCGImage:imageRef]]; 
CGImageRelease(imageRef);

To crop retina images while keeping the same scale and orientation, use the following method in a UIImage category (iOS 4.0 and above):

- (UIImage *)crop:(CGRect)rect {
	if (self.scale > 1.0f) {
		rect = CGRectMake(rect.origin.x * self.scale,
						  rect.origin.y * self.scale,
						  rect.size.width * self.scale,
						  rect.size.height * self.scale);
	}
	
    CGImageRef imageRef = CGImageCreateWithImageInRect(self.CGImage, rect);
    UIImage *result = [UIImage imageWithCGImage:imageRef scale:self.scale orientation:self.imageOrientation];
    CGImageRelease(imageRef);
    return result;
}

You can make a UIImage category and use it wherever you need. Based on HitScans response and comments bellow it.

@implementation UIImage (Crop)

- (UIImage *)crop:(CGRect)rect {

    rect = CGRectMake(rect.origin.x*self.scale, 
                      rect.origin.y*self.scale, 
                      rect.size.width*self.scale, 
                      rect.size.height*self.scale);       
    
    CGImageRef imageRef = CGImageCreateWithImageInRect([self CGImage], rect);
    UIImage *result = [UIImage imageWithCGImage:imageRef 
                                          scale:self.scale 
                                    orientation:self.imageOrientation]; 
    CGImageRelease(imageRef);
    return result;
}

@end

You can use it this way:

UIImage *imageToCrop = <yourImageToCrop>;
CGRect cropRect = <areaYouWantToCrop>;   

//for example
//CGRectMake(0, 40, 320, 100);
    
UIImage *croppedImage = [imageToCrop crop:cropRect];

Swift 3 version

func cropImage(imageToCrop:UIImage, toRect rect:CGRect) -> UIImage{
    
    let imageRef:CGImage = imageToCrop.cgImage!.cropping(to: rect)!
    let cropped:UIImage = UIImage(cgImage:imageRef)
    return cropped
}


let imageTop:UIImage  = UIImage(named:"one.jpg")! // add validation

with help of this bridge function CGRectMake -> CGRect (credits to this answer answered by @rob mayoff):

 func CGRectMake(_ x: CGFloat, _ y: CGFloat, _ width: CGFloat, _ height: CGFloat) -> CGRect {
    return CGRect(x: x, y: y, width: width, height: height)
}

The usage is:

if var image:UIImage  = UIImage(named:"one.jpg"){
   let  croppedImage = cropImage(imageToCrop: image, toRect: CGRectMake(
        image.size.width/4,
        0,
        image.size.width/2,
        image.size.height)
    )
}

Output:

Here is my UIImage crop implementation which obeys the imageOrientation property. All orientations were thoroughly tested.

inline double rad(double deg)
{
	return deg / 180.0 * M_PI;
}

UIImage* UIImageCrop(UIImage* img, CGRect rect)
{
	CGAffineTransform rectTransform;
	switch (img.imageOrientation)
	{
		case UIImageOrientationLeft:
			rectTransform = CGAffineTransformTranslate(CGAffineTransformMakeRotation(rad(90)), 0, -img.size.height);
			break;
		case UIImageOrientationRight:
			rectTransform = CGAffineTransformTranslate(CGAffineTransformMakeRotation(rad(-90)), -img.size.width, 0);
			break;
		case UIImageOrientationDown:
			rectTransform = CGAffineTransformTranslate(CGAffineTransformMakeRotation(rad(-180)), -img.size.width, -img.size.height);
			break;
		default:
			rectTransform = CGAffineTransformIdentity;
	};
	rectTransform = CGAffineTransformScale(rectTransform, img.scale, img.scale);
	
	CGImageRef imageRef = CGImageCreateWithImageInRect([img CGImage], CGRectApplyAffineTransform(rect, rectTransform));
	UIImage *result = [UIImage imageWithCGImage:imageRef scale:img.scale orientation:img.imageOrientation];
	CGImageRelease(imageRef);
	return result;
}

Heads up: all these answers assume a CGImage-backed image object.

image.CGImage can return nil, if the UIImage is backed by a CIImage, which would be the case if you created this image using a CIFilter.

In that case, you might have to draw the image in a new context, and return that image (slow).

UIImage* crop(UIImage *image, rect) {
    UIGraphicsBeginImageContextWithOptions(rect.size, false, [image scale]);
    [image drawAtPoint:CGPointMake(-rect.origin.x, -rect.origin.y)];
    cropped_image = UIGraphicsGetImageFromCurrentImageContext();
    UIGraphicsEndImageContext();
    return cropped_image;
}

None of the answers here handle all of the scale and rotation issues 100% correctly. Here's a synthesis of everything said so far, up-to-date as of iOS7/8. It's meant to be included as a method in a category on UIImage.

- (UIImage *)croppedImageInRect:(CGRect)rect
{
    double (^rad)(double) = ^(double deg) {
        return deg / 180.0 * M_PI;
    };

    CGAffineTransform rectTransform;
    switch (self.imageOrientation) {
        case UIImageOrientationLeft:
            rectTransform = CGAffineTransformTranslate(CGAffineTransformMakeRotation(rad(90)), 0, -self.size.height);
            break;
        case UIImageOrientationRight:
            rectTransform = CGAffineTransformTranslate(CGAffineTransformMakeRotation(rad(-90)), -self.size.width, 0);
            break;
        case UIImageOrientationDown:
            rectTransform = CGAffineTransformTranslate(CGAffineTransformMakeRotation(rad(-180)), -self.size.width, -self.size.height);
            break;
        default:
            rectTransform = CGAffineTransformIdentity;
    };
    rectTransform = CGAffineTransformScale(rectTransform, self.scale, self.scale);

    CGImageRef imageRef = CGImageCreateWithImageInRect([self CGImage], CGRectApplyAffineTransform(rect, rectTransform));
    UIImage *result = [UIImage imageWithCGImage:imageRef scale:self.scale orientation:self.imageOrientation];
    CGImageRelease(imageRef);

    return result;
}

Swift version of awolf's answer, which worked for me:

public extension UIImage {
    func croppedImage(inRect rect: CGRect) -> UIImage {
        let rad: (Double) -> CGFloat = { deg in
            return CGFloat(deg / 180.0 * .pi)
        }
        var rectTransform: CGAffineTransform
        switch imageOrientation {
        case .left:
            let rotation = CGAffineTransform(rotationAngle: rad(90))
            rectTransform = rotation.translatedBy(x: 0, y: -size.height)
        case .right:
            let rotation = CGAffineTransform(rotationAngle: rad(-90))
            rectTransform = rotation.translatedBy(x: -size.width, y: 0)
        case .down:
            let rotation = CGAffineTransform(rotationAngle: rad(-180))
            rectTransform = rotation.translatedBy(x: -size.width, y: -size.height)
        default:
            rectTransform = .identity
        }
        rectTransform = rectTransform.scaledBy(x: scale, y: scale)
        let transformedRect = rect.applying(rectTransform)
        let imageRef = cgImage!.cropping(to: transformedRect)!
        let result = UIImage(cgImage: imageRef, scale: scale, orientation: imageOrientation)
        return result
    }
}

CGSize size = [originalImage size];
int padding = 20;
int pictureSize = 300;
int startCroppingPosition = 100;
if (size.height > size.width) {
	pictureSize = size.width - (2.0 * padding);
	startCroppingPosition = (size.height - pictureSize) / 2.0; 
} else {
	pictureSize = size.height - (2.0 * padding);
	startCroppingPosition = (size.width - pictureSize) / 2.0;
}
// WTF: Don't forget that the CGImageCreateWithImageInRect believes that 
// the image is 180 rotated, so x and y are inverted, same for height and width.
CGRect cropRect = CGRectMake(startCroppingPosition, padding, pictureSize, pictureSize);
CGImageRef imageRef = CGImageCreateWithImageInRect([originalImage CGImage], cropRect);
UIImage *newImage = [UIImage imageWithCGImage:imageRef scale:1.0 orientation:originalImage.imageOrientation];
[m_photoView setImage:newImage];
CGImageRelease(imageRef);

Most of the responses I've seen only deals with a position of (0, 0) for (x, y). Ok that's one case but I'd like my cropping operation to be centered. What took me a while to figure out is the line following the WTF comment.

Let's take the case of an image captured with a portrait orientation:

1. The original image height is higher than its width (Woo, no surprise so far!)
2. The image that the CGImageCreateWithImageInRect method imagines in its own world is not really a portrait though but a landscape (That is also why if you don't use the orientation argument in the imageWithCGImage constructor, it will show up as 180 rotated).
3. So, you should kind of imagine that it is a landscape, the (0, 0) position being the top right corner of the image.

Hope it makes sense! If it does not, try different values you'll see that the logic is inverted when it comes to choosing the right x, y, width, and height for your cropRect.

swift3

extension UIImage {
	func crop(rect: CGRect) -> UIImage? {
		var scaledRect = rect
		scaledRect.origin.x *= scale
		scaledRect.origin.y *= scale
		scaledRect.size.width *= scale
		scaledRect.size.height *= scale
		guard let imageRef: CGImage = cgImage?.cropping(to: scaledRect) else {
			return nil
		}
		return UIImage(cgImage: imageRef, scale: scale, orientation: imageOrientation)
	}
}

Swift Extension

extension UIImage {
    func crop(var rect: CGRect) -> UIImage {
        rect.origin.x*=self.scale
        rect.origin.y*=self.scale
        rect.size.width*=self.scale
        rect.size.height*=self.scale

        let imageRef = CGImageCreateWithImageInRect(self.CGImage, rect)
        let image = UIImage(CGImage: imageRef, scale: self.scale, orientation: self.imageOrientation)!
        return image
    }
}

Best solution for cropping an UIImage in Swift, in term of precision, pixels scaling ...:

private func squareCropImageToSideLength(let sourceImage: UIImage,
    let sideLength: CGFloat) -> UIImage {
        // input size comes from image
        let inputSize: CGSize = sourceImage.size
        
        // round up side length to avoid fractional output size
        let sideLength: CGFloat = ceil(sideLength)
        
        // output size has sideLength for both dimensions
        let outputSize: CGSize = CGSizeMake(sideLength, sideLength)
        
        // calculate scale so that smaller dimension fits sideLength
        let scale: CGFloat = max(sideLength / inputSize.width,
            sideLength / inputSize.height)
        
        // scaling the image with this scale results in this output size
        let scaledInputSize: CGSize = CGSizeMake(inputSize.width * scale,
            inputSize.height * scale)
        
        // determine point in center of "canvas"
        let center: CGPoint = CGPointMake(outputSize.width/2.0,
            outputSize.height/2.0)
        
        // calculate drawing rect relative to output Size
        let outputRect: CGRect = CGRectMake(center.x - scaledInputSize.width/2.0,
            center.y - scaledInputSize.height/2.0,
            scaledInputSize.width,
            scaledInputSize.height)
        
        // begin a new bitmap context, scale 0 takes display scale
        UIGraphicsBeginImageContextWithOptions(outputSize, true, 0)
        
        // optional: set the interpolation quality.
        // For this you need to grab the underlying CGContext
        let ctx: CGContextRef = UIGraphicsGetCurrentContext()
        CGContextSetInterpolationQuality(ctx, kCGInterpolationHigh)
        
        // draw the source image into the calculated rect
        sourceImage.drawInRect(outputRect)
        
        // create new image from bitmap context
        let outImage: UIImage = UIGraphicsGetImageFromCurrentImageContext()
        
        // clean up
        UIGraphicsEndImageContext()
        
        // pass back new image
        return outImage
}

Instructions used to call this function:

let image: UIImage = UIImage(named: "Image.jpg")!
let squareImage: UIImage = self.squareCropImageToSideLength(image, sideLength: 320)
self.myUIImageView.image = squareImage

Note: the initial source code inspiration written in Objective-C has been found on "Cocoanetics" blog.

Below code snippet might help.

import UIKit

extension UIImage {
    func cropImage(toRect rect: CGRect) -> UIImage? {
        if let imageRef = self.cgImage?.cropping(to: rect) {
            return UIImage(cgImage: imageRef)
        }
        return nil
    }
}

Looks a little bit strange but works great and takes into consideration image orientation:

var image:UIImage = ...

let img = CIImage(image: image)!.imageByCroppingToRect(rect)
image = UIImage(CIImage: img, scale: 1, orientation: image.imageOrientation)

- (UIImage *)getSubImage:(CGRect) rect{
    CGImageRef subImageRef = CGImageCreateWithImageInRect(self.CGImage, rect);
    CGRect smallBounds = CGRectMake(rect.origin.x, rect.origin.y, CGImageGetWidth(subImageRef), CGImageGetHeight(subImageRef));
    
    UIGraphicsBeginImageContext(smallBounds.size);
    CGContextRef context = UIGraphicsGetCurrentContext();
    CGContextDrawImage(context, smallBounds, subImageRef);
    UIImage* smallImg = [UIImage imageWithCGImage:subImageRef];
    UIGraphicsEndImageContext();
    
    return smallImg;
}

 (UIImage *)squareImageWithImage:(UIImage *)image scaledToSize:(CGSize)newSize {
    double ratio;
    double delta;
    CGPoint offset;
 
    //make a new square size, that is the resized imaged width
    CGSize sz = CGSizeMake(newSize.width, newSize.width);
 
    //figure out if the picture is landscape or portrait, then
    //calculate scale factor and offset
    if (image.size.width > image.size.height) {
        ratio = newSize.width / image.size.width;
        delta = (ratio*image.size.width - ratio*image.size.height);
        offset = CGPointMake(delta/2, 0);
    } else {
        ratio = newSize.width / image.size.height;
        delta = (ratio*image.size.height - ratio*image.size.width);
        offset = CGPointMake(0, delta/2);
    }
 
    //make the final clipping rect based on the calculated values
    CGRect clipRect = CGRectMake(-offset.x, -offset.y,
                                 (ratio * image.size.width) + delta,
                                 (ratio * image.size.height) + delta);
 
 
    //start a new context, with scale factor 0.0 so retina displays get
    //high quality image
    if ([[UIScreen mainScreen] respondsToSelector:@selector(scale)]) {
        UIGraphicsBeginImageContextWithOptions(sz, YES, 0.0);
    } else {
        UIGraphicsBeginImageContext(sz);
    }
    UIRectClip(clipRect);
    [image drawInRect:clipRect];
    UIImage *newImage = UIGraphicsGetImageFromCurrentImageContext();
    UIGraphicsEndImageContext();
 
    return newImage;
}

On iOS9.2SDK ,I use below method to convert frame from UIView to UIImage

-(UIImage *)getNeedImageFrom:(UIImage*)image cropRect:(CGRect)rect
{
  CGSize cropSize = rect.size;
  CGFloat widthScale = image.size.width/self.imageViewOriginal.bounds.size.width;
  CGFloat heightScale = image.size.height/self.imageViewOriginal.bounds.size.height;
  cropSize = CGSizeMake(rect.size.width*widthScale, 
              rect.size.height*heightScale);
  CGPoint pointCrop = CGPointMake(rect.origin.x*widthScale,
             rect.origin.y*heightScale);
  rect = CGRectMake(pointCrop.x, pointCrop.y, cropSize.width, cropSize.height);
  CGImageRef subImage = CGImageCreateWithImageInRect(image.CGImage, rect);
  UIImage *croppedImage = [UIImage imageWithCGImage:subImage];
  CGImageRelease(subImage);

  return croppedImage;
}

Swift 2.0 Update (CIImage compatibility)

Expanding off of Maxim's Answer but works if your image is CIImage based, as well.

public extension UIImage {
    func imageByCroppingToRect(rect: CGRect) -> UIImage? {
        if let image = CGImageCreateWithImageInRect(self.CGImage, rect) {
            return UIImage(CGImage: image)
        } else if let image = (self.CIImage)?.imageByCroppingToRect(rect) {
            return UIImage(CIImage: image)
        }
       return nil
   }
}

Here's an updated Swift 3 version based on Noodles answer

func cropping(to rect: CGRect) -> UIImage? {

    if let cgCrop = cgImage?.cropping(to: rect) {
        return UIImage(cgImage: cgCrop)
    }
    else if let ciCrop = ciImage?.cropping(to: rect) {
        return UIImage(ciImage: ciCrop)
    }
    
    return nil
}

Follow Answer of @Arne. I Just fixing to Category function. put it in Category of UIImage.

-(UIImage*)cropImage:(CGRect)rect{
    
    UIGraphicsBeginImageContextWithOptions(rect.size, false, [self scale]);
    [self drawAtPoint:CGPointMake(-rect.origin.x, -rect.origin.y)];
    UIImage* cropped_image = UIGraphicsGetImageFromCurrentImageContext();
    UIGraphicsEndImageContext();
    return cropped_image;
}

Swift 5:

extension UIImage {
	func cropped(rect: CGRect) -> UIImage? {
        guard let cgImage = cgImage else { return nil }
        
        UIGraphicsBeginImageContextWithOptions(rect.size, false, 0)
        let context = UIGraphicsGetCurrentContext()

        context?.translateBy(x: 0.0, y: self.size.height)
        context?.scaleBy(x: 1.0, y: -1.0)
        context?.draw(cgImage, in: CGRect(x: rect.minX, y: rect.minY, width: self.size.width, height: self.size.height), byTiling: false)


        let croppedImage = UIGraphicsGetImageFromCurrentImageContext()
        UIGraphicsEndImageContext()

        return croppedImage
    }
}

I wasn't satisfied with other solutions because they either draw several time (using more power than necessary) or have problems with orientation. Here is what I used for a scaled square croppedImage from a UIImage * image.

CGFloat minimumSide = fminf(image.size.width, image.size.height);
CGFloat finalSquareSize = 600.;

//create new drawing context for right size
CGRect rect = CGRectMake(0, 0, finalSquareSize, finalSquareSize);
CGFloat scalingRatio = 640.0/minimumSide;
UIGraphicsBeginImageContext(rect.size);
    
//draw
[image drawInRect:CGRectMake((minimumSide - photo.size.width)*scalingRatio/2., (minimumSide - photo.size.height)*scalingRatio/2., photo.size.width*scalingRatio, photo.size.height*scalingRatio)];

UIImage *croppedImage = UIGraphicsGetImageFromCurrentImageContext();

UIGraphicsEndImageContext();

I use the method below.

  -(UIImage *)getNeedImageFrom:(UIImage*)image cropRect:(CGRect)rect
  {
    CGSize cropSize = rect.size;
    CGFloat widthScale =  
    image.size.width/self.imageViewOriginal.bounds.size.width;
    CGFloat heightScale = 
    image.size.height/self.imageViewOriginal.bounds.size.height;
    cropSize = CGSizeMake(rect.size.width*widthScale,  
    rect.size.height*heightScale);
    CGPoint  pointCrop = CGPointMake(rect.origin.x*widthScale, 
    rect.origin.y*heightScale);
    rect = CGRectMake(pointCrop.x, pointCrop.y, cropSize.width, 
    cropSize.height);
    CGImageRef subImage = CGImageCreateWithImageInRect(image.CGImage, rect);
    UIImage *croppedImage = [UIImage imageWithCGImage:subImage];
    CGImageRelease(subImage);
    return croppedImage;

}

Look at https://github.com/vvbogdan/BVCropPhoto

(UIImage *)croppedImage {
CGFloat scale = self.sourceImage.size.width / self.scrollView.contentSize.width;
UIImage *finalImage = nil;
CGRect targetFrame = CGRectMake((self.scrollView.contentInset.left + self.scrollView.contentOffset.x) * scale,
(self.scrollView.contentInset.top + self.scrollView.contentOffset.y) * scale,
self.cropSize.width * scale,
self.cropSize.height * scale);
CGImageRef contextImage = CGImageCreateWithImageInRect([[self imageWithRotation:self.sourceImage] CGImage], targetFrame);
if (contextImage != NULL) {
finalImage = [UIImage imageWithCGImage:contextImage
scale:self.sourceImage.scale
orientation:UIImageOrientationUp];
  CGImageRelease(contextImage);

}
return finalImage;
}


(UIImage *)imageWithRotation:(UIImage *)image {
if (image.imageOrientation == UIImageOrientationUp) return image;
CGAffineTransform transform = CGAffineTransformIdentity;
switch (image.imageOrientation) {
case UIImageOrientationDown:
case UIImageOrientationDownMirrored:
transform = CGAffineTransformTranslate(transform, image.size.width, image.size.height);
transform = CGAffineTransformRotate(transform, M_PI);
break;
  case UIImageOrientationLeft:
  case UIImageOrientationLeftMirrored:
      transform = CGAffineTransformTranslate(transform, image.size.width, 0);
      transform = CGAffineTransformRotate(transform, M_PI_2);
      break;

  case UIImageOrientationRight:
  case UIImageOrientationRightMirrored:
      transform = CGAffineTransformTranslate(transform, 0, image.size.height);
      transform = CGAffineTransformRotate(transform, -M_PI_2);
      break;
  case UIImageOrientationUp:
  case UIImageOrientationUpMirrored:
      break;

}
switch (image.imageOrientation) {
case UIImageOrientationUpMirrored:
case UIImageOrientationDownMirrored:
transform = CGAffineTransformTranslate(transform, image.size.width, 0);
transform = CGAffineTransformScale(transform, -1, 1);
break;
  case UIImageOrientationLeftMirrored:
  case UIImageOrientationRightMirrored:
      transform = CGAffineTransformTranslate(transform, image.size.height, 0);
      transform = CGAffineTransformScale(transform, -1, 1);
      break;
  case UIImageOrientationUp:
  case UIImageOrientationDown:
  case UIImageOrientationLeft:
  case UIImageOrientationRight:
      break;

}
// Now we draw the underlying CGImage into a new context, applying the transform
// calculated above.
CGContextRef ctx = CGBitmapContextCreate(NULL, image.size.width, image.size.height,
CGImageGetBitsPerComponent(image.CGImage), 0,
CGImageGetColorSpace(image.CGImage),
CGImageGetBitmapInfo(image.CGImage));
CGContextConcatCTM(ctx, transform);
switch (image.imageOrientation) {
case UIImageOrientationLeft:
case UIImageOrientationLeftMirrored:
case UIImageOrientationRight:
case UIImageOrientationRightMirrored:
// Grr...
CGContextDrawImage(ctx, CGRectMake(0, 0, image.size.height, image.size.width), image.CGImage);
break;
  default:
      CGContextDrawImage(ctx, CGRectMake(0, 0, image.size.width, image.size.height), image.CGImage);
      break;

}
// And now we just create a new UIImage from the drawing context
CGImageRef cgimg = CGBitmapContextCreateImage(ctx);
UIImage *img = [UIImage imageWithCGImage:cgimg];
CGContextRelease(ctx);
CGImageRelease(cgimg);
return img;


}

Swift 5.0 update

public extension UIImage {
    func cropped(rect: CGRect) -> UIImage? {
        if let image = self.cgImage!.cropping(to: rect) {
            return UIImage(cgImage: image)
        } else if let image = (self.ciImage)?.cropped(to: rect) {
            return UIImage(ciImage: image)
        }
       return nil
   }
}