浅谈Visual C#进行图像处理（读取、保存以及对像素的访问）

这里之所以说“浅谈”是因为我这里只是简单的介绍如何使用visual c#进行图像的读入、保存以及对像素的访问。而不涉及太多的算法。

一、读取图像

在visual c#中我们可以使用一个picture box控件来显示图片，如下：

private void btnopenimage_click(object sender, eventargs e)
{
    openfiledialog ofd = new openfiledialog();
    ofd.filter = "bmp files(*.bmp)|*.bmp|jpg files(*.jpg;*.jpeg)|*.jpg;*.jpeg|all files(*.*)|*.*";
    ofd.checkfileexists = true;
    ofd.checkpathexists = true;
    if (ofd.showdialog() == dialogresult.ok)
    {
        //pbxshowimage.imagelocation = ofd.filename;
        bmp = new bitmap(ofd.filename);
        if (bmp==null)
        {
            messagebox.show("加载图片失败!", "错误");
            return;
        }
        pbxshowimage.image = bmp;
        ofd.dispose();
    }
}

其中bmp为类的一个对象：private bitmap bmp=null;
在使用bitmap类和bitmapdata类之前，需要使用using system.drawing.imaging;

二、保存图像

private void btnsaveimage_click(object sender, eventargs e)
{
    if (bmp == null) return;
    savefiledialog sfd = new savefiledialog();
    sfd.filter = "bmp files(*.bmp)|*.bmp|jpg files(*.jpg;*.jpeg)|*.jpg;*.jpeg|all files(*.*)|*.*";
    if (sfd.showdialog() == dialogresult.ok)
    {
        pbxshowimage.image.save(sfd.filename);
        messagebox.show("保存成功!","提示");
        sfd.dispose();
    }
}

三、对像素的访问

我们可以来建立一个graybitmapdata类来做相关的处理。整个类的程序如下：

using system;
using system.collections.generic;
using system.linq;
using system.text;
using system.drawing;
using system.drawing.imaging;
using system.windows.forms;
namespace imageelf
{
    class graybitmapdata
    {
        public byte[,] data;//保存像素矩阵
        public int width;//图像的宽度
        public int height;//图像的高度
        public graybitmapdata()
        {
            this.width = 0;
            this.height = 0;
            this.data = null;
        }
        public graybitmapdata(bitmap bmp)
        {
            bitmapdata bmpdata = bmp.lockbits(new rectangle(0, 0, bmp.width, bmp.height), imagelockmode.readonly, pixelformat.format24bpprgb);
            this.width = bmpdata.width;
            this.height = bmpdata.height;
            data = new byte[height, width];
            unsafe
            {
                byte* ptr = (byte*)bmpdata.scan0.topointer();
                for (int i = 0; i < height; i++)
                {
                    for (int j = 0; j < width; j++)
                    {
    //将24位的rgb彩色图转换为灰度图
                        int temp = (int)(0.114 * (*ptr++)) + (int)(0.587 * (*ptr++))+(int)(0.299 * (*ptr++));
                        data[i, j] = (byte)temp;
                    }
                    ptr += bmpdata.stride - width * 3;//指针加上填充的空白空间
                }
            }
            bmp.unlockbits(bmpdata);
        }
        public graybitmapdata(string path)
            : this(new bitmap(path))
        {
        }
        public bitmap tobitmap()
        {
            bitmap bmp=new bitmap(width,height,pixelformat.format24bpprgb);
            bitmapdata bmpdata=bmp.lockbits(new rectangle(0,0,width,height),imagelockmode.writeonly,pixelformat.format24bpprgb);
            unsafe
            {
                byte* ptr=(byte*)bmpdata.scan0.topointer();
                for(int i=0;i<height;i++)
                {
                    for(int j=0;j<width;j++)
                    {
                        *(ptr++)=data[i,j];
                        *(ptr++)=data[i,j];
                        *(ptr++)=data[i,j];
                    }
                    ptr+=bmpdata.stride-width*3;
                }
            }
            bmp.unlockbits(bmpdata);
            return bmp;
        }
        public void showimage(picturebox pbx)
        {
            bitmap b = this.tobitmap();
            pbx.image = b;
            //b.dispose();
        }
        public void saveimage(string path)
        {
            bitmap b=tobitmap();
            b.save(path);
            //b.dispose();
        }
//均值滤波
        public void averagefilter(int windowsize)
        {
            if (windowsize % 2 == 0)
            {
                return;
            }
            for (int i = 0; i < height; i++)
            {
                for (int j = 0; j < width; j++)
                {
                    int sum = 0;
                    for (int g = -(windowsize - 1) / 2; g <= (windowsize - 1) / 2; g++)
                    {
                        for (int k = -(windowsize - 1) / 2; k <= (windowsize - 1) / 2; k++)
                        {
                            int a = i + g, b = j + k;
                            if (a < 0) a = 0;
                            if (a > height - 1) a = height - 1;
                            if (b < 0) b = 0;
                            if (b > width - 1) b = width - 1;
                            sum += data[a, b];
                        }
                    }
                    data[i,j]=(byte)(sum/(windowsize*windowsize));
                }
            }
        }
//中值滤波
        public void midfilter(int windowsize)
        {
            if (windowsize % 2 == 0)
            {
                return;
            }
            int[] temp = new int[windowsize * windowsize];
            byte[,] newdata = new byte[height, width];
            for (int i = 0; i < height; i++)
            {
                for (int j = 0; j < width; j++)
                {
                    int n = 0;
                    for (int g = -(windowsize - 1) / 2; g <= (windowsize - 1) / 2; g++)
                    {
                        for (int k = -(windowsize - 1) / 2; k <= (windowsize - 1) / 2; k++)
                        {
                            int a = i + g, b = j + k;
                            if (a < 0) a = 0;
                            if (a > height - 1) a = height - 1;
                            if (b < 0) b = 0;
                            if (b > width - 1) b = width - 1;
                            temp[n++]= data[a, b];
                        }
                    }
                    newdata[i, j] = getmidvalue(temp,windowsize*windowsize);
                }
            }
            for (int i = 0; i < height; i++)
            {
                for (int j = 0; j < width; j++)
                {
                    data[i, j] = newdata[i, j];
                }
            }
        }
//获得一个向量的中值
        private byte getmidvalue(int[] t, int length)
        {
            int temp = 0;
            for (int i = 0; i < length - 2; i++)
            {
                for (int j = i + 1; j < length - 1; j++)
                {
                    if (t[i] > t[j])
                    {
                        temp = t[i];
                        t[i] = t[j];
                        t[j] = temp;
                    }
                }
            }
            return (byte)t[(length - 1) / 2];
        }
//一种新的滤波方法，是亮的更亮、暗的更暗
        public void newfilter(int windowsize)
        {
            if (windowsize % 2 == 0)
            {
                return;
            }
            for (int i = 0; i < height; i++)
            {
                for (int j = 0; j < width; j++)
                {
                    int sum = 0;
                    for (int g = -(windowsize - 1) / 2; g <= (windowsize - 1) / 2; g++)
                    {
                        for (int k = -(windowsize - 1) / 2; k <= (windowsize - 1) / 2; k++)
                        {
                            int a = i + g, b = j + k;
                            if (a < 0) a = 0;
                            if (a > height - 1) a = height - 1;
                            if (b < 0) b = 0;
                            if (b > width - 1) b = width - 1;
                            sum += data[a, b];
                        }
                    }
                    double avg = (sum+0.0) / (windowsize * windowsize);
                    if (avg / 255 < 0.5)
                    {
                        data[i, j] = (byte)(2 * avg / 255 * data[i, j]);
                    }
                    else
                    {
                        data[i,j]=(byte)((1-2*(1-avg/255.0)*(1-data[i,j]/255.0))*255);
                    }
                }
            }
        }
//直方图均衡
        public void histequal()
        {
            double[] num = new double[256] ;
            for(int i=0;i<256;i++) num[i]=0;
            for (int i = 0; i < height; i++)
            {
                for (int j = 0; j < width; j++)
                {
                    num[data[i, j]]++;
                }
            }
            double[] newgray = new double[256];
            double n = 0;
            for (int i = 0; i < 256; i++)
            {
                n += num[i];
                newgray[i] = n * 255 / (height * width);
            }
            for (int i = 0; i < height; i++)
            {
                for (int j = 0; j < width; j++)
                {
                    data[i,j]=(byte)newgray[data[i,j]];
                }
            }
        }
}
}

在graybitmapdata类中，只要我们对一个二维数组data进行一系列的操作就是对图片的操作处理。在窗口上，我们可以使用
一个按钮来做各种调用：

//均值滤波
private void btnavgfilter_click(object sender, eventargs e)
{
    if (bmp == null) return;
    graybitmapdata gbmp = new graybitmapdata(bmp);
    gbmp.averagefilter(3);
    gbmp.showimage(pbxshowimage);
}
//转换为灰度图
private void btntogray_click(object sender, eventargs e)
{
    if (bmp == null) return;
    graybitmapdata gbmp = new graybitmapdata(bmp);
    gbmp.showimage(pbxshowimage);
}

四、总结

在visual c#中对图像进行处理或访问，需要先建立一个bitmap对象，然后通过其lockbits方法来获得一个bitmapdata类的对象，然后通过获得其像素数据的首地址来对bitmap对象的像素数据进行操作。当然，一种简单但是速度慢的方法是用bitmap类的getpixel和setpixel方法。其中bitmapdata类的stride属性为每行像素所占的字节。