Implementing Image Scaling and Cropping in .NET Core - Based on ImageSharp

Preface

(Suddenly realized I haven't updated in a while

Recently while working on my blog, I needed to implement a random image feature similar to Lorempixel and LoremPicsum. I already had the images, and I also needed a feature to get an image based on the input width and height. Since I had used the ImageSharp library when processing images before, I immediately thought of using it this time.

Analyzing Requirements

The images in the image library are basically wallpapers I collected before, with uneven sizes, some landscape and some portrait

Then wrapping it as an API that only requires inputting width and height, it can randomly pick an image and then scale or crop it

My approach is:

• For landscape images, adjust the height to match the input height, and adjust the width proportionally
• For portrait images, adjust the width to match the input height, and adjust the height proportionally

Then select the middle portion for cropping

Of course there's also a special case, when the input width and height exceed the original height and width of the image; in this case we can only break the original aspect ratio of the image and force a stretch~

PS: I originally wanted to draw a diagram to express the idea, but I couldn't find a handy drawing tool (Draw.io: don't look at me)

Implementation

First, read the image

Very simple, just pass in the image path

Of course you can also read it via a stream

using var image = Image.Load("imagePath");

Then, based on the various size cases of the image, perform scaling and cropping

A lot of the blog posts I found online use code similar to image.Resize and image.Crop, but this is the old version of ImageSharp code

In the new version everything is placed in image.Mutate, and you pass in a lambda expression for whatever operation you need

(A bit like ORM operations)

For example, scaling looks like this

image.Mutate(a => a.Resize(newWidth, newHeight));

Cropping looks like this

image.Mutate(a => a.Crop(new Rectangle(x, y, width, height)));

Complete Code

The functionality is very simple; the complete code is here

void GetImage(string imagePath, int width, int height) {
    using var image = Image.Load(imagePath);
    Rectangle cropRect;
    int newWidth;
    int newHeight;
    
    // 横屏图片
    if (image.Width > image.Height) {
        if (width > image.Width) {
            newWidth = width;
            newHeight = height;
        }
        else {
            newHeight = height;
            newWidth = image.Width / image.Height * newHeight;
        }

        cropRect = new Rectangle((newWidth - width) / 2, 0, width, height);
    }
    // 竖屏图片
    else {
        if (height > image.Height) {
            newWidth = width;
            newHeight = height;
        }
        else {
            newWidth = width;
            newHeight = newWidth * image.Height / image.Width;
        }

        cropRect = new Rectangle(0, (newHeight - height) / 2, width, height);
    }

    image.Mutate(a => a.Resize(newWidth, newHeight));
    image.Mutate(a => a.Crop(cropRect));
    image.SaveAsPng("output.png");
}

Follow-up will be updated in my StarBlog development notes series~

References

• Random image website: https://picsum.photos/
• How to use ImageSharp for image generation in .NET Core: https://www.cnblogs.com/niwan/p/11126239.html
• https://stackoverflow.com/questions/63639644/how-load-an-image-from-disk-and-save-to-a-stream-using-imagesharp-while-preservi