网络运营公司的经营范围关键词排名优化易下拉软件

本项目基于Emgu.CV（C#下OpenCv的封装）开发的，编译器最新版Vs2022，编译环境x86

直接看效果图

1.主页面

2.我们先看手写的方式：

点击确认就到主界面，如下 ：

点击自动适配-，再点击生成：

放大看

 

点击保存即可，生成透明电子签名图片。

3.在单色背景下手写名字，导入图片生成

先点击 选择图像 按钮，然后选择图像，点击自动适配，点击Canny算法生成，最后点击生成，如下：

双击右边第一张放大：

选择需要的区域点击保存即可。

4.算法介绍：

4.1自动适配

自动适配指根据灰度值像素选取Canny算子的上下阈值,主要目的是能快速自动适配Canny上下阈值，这里主要介绍以中位数或者平均值，然后以Sigma为临界值选取一个区间，如下：

        /// <summary>/// 自动适配/// </summary>/// <param name="sender"></param>/// <param name="e"></param>private void button4_Click(object sender, EventArgs e){if (pictureBox1.Image == null){MessageBox.Show("请选择图像");return;}imgShow?.Dispose();gray?.Dispose();imgShow = new Image<Bgra, byte>(new Bitmap(pictureBox1.Image));gray = imgShow.Convert<Gray, byte>();double median = radioButton1.Checked ? CalculateMedian(gray) : CalculateAvg(gray);double sigma = (float)numericUpDown1.Value;double lower = Math.Max(0, (1.0 - sigma) * median);double upper = Math.Min(255, (1.0 + sigma) * median);trackBar1.Value = (int)lower;label1.Text = trackBar1.Value.ToString();trackBar2.Value = (int)upper;label3.Text = trackBar2.Value.ToString();}

 4.2Canny生成

Canny算子边缘检测具体不多介绍，这里目的是为了找到字体边缘，然后在图像上填充，方便我们能确定这个边缘是否能够准确找到，如果大都填充到字体上，说明边缘检测不错，找到这些边缘，然后计算它们的平均R、G、B：

        /// <summary>/// Canny生成/// </summary>/// <param name="sender"></param>/// <param name="e"></param>private void button6_Click(object sender, EventArgs e){if (gray == null || imgShow == null){MessageBox.Show("请先选择自动适配");return;}// Canny算子边缘检测using Image<Gray, byte> edges = gray.Canny(trackBar1.Value, trackBar2.Value);using VectorOfVectorOfPoint contours = new VectorOfVectorOfPoint();using Mat hier = new Mat();CvInvoke.FindContours(edges, contours, hier, RetrType.List, ChainApproxMethod.ChainApproxSimple);// 掩码using Image<Gray, byte> mask = new Image<Gray, byte>(gray.Size);if (contours.Size > 10000){MessageBox.Show("边缘太多，会非常耗时，请重先调节参数");return;}// 在imgShow图像上填充边缘边框for (int i = 0; i < contours.Size; i++){var contour = contours[i];CvInvoke.DrawContours(imgShow, contours, i, new MCvScalar(0, 0, 255), 2);CvInvoke.DrawContours(mask, contours, i, new MCvScalar(255), thickness: -1);}this.pictureBox3.Image?.Dispose();this.pictureBox3.Image = imgShow.Bitmap;using Image<Gray, byte> maskedImage = new Image<Gray, byte>(gray.Size);CvInvoke.BitwiseAnd(gray, gray, maskedImage, mask);MCvScalar average = CvInvoke.Mean(maskedImage, mask);double blue = average.V0;double green = average.V1;double red = average.V2;trackBar3.Value = (int)Math.Round(red, 0);trackBar4.Value = (int)Math.Round(green, 0);trackBar5.Value = (int)Math.Round(blue, 0);label9.Text = trackBar3.Value.ToString();label10.Text = trackBar4.Value.ToString();label12.Text = trackBar5.Value.ToString();}

4.3生成

在生成中，我们会根据R、G、B计算出掩码，然后将掩码中选取的颜色的alpha通道设置为0，即为透明。

        /// <summary>/// 生成/// </summary>/// <param name="sender"></param>/// <param name="e"></param>private void button2_Click(object sender, EventArgs e){if (pictureBox1.Image == null){MessageBox.Show("请选择图片");return;}img?.Dispose();img = new Image<Bgra, byte>(new Bitmap(pictureBox1.Image));// 掩码using Image<Gray, byte> mask2 = img.InRange(new Bgra(trackBar5.Value, trackBar4.Value, trackBar3.Value, 0), new Bgra(255, 255, 255, 255));// 将掩码中选取的颜色的alpha通道设置为0img.SetValue(new Bgra(0, 0, 0, 0), mask2);this.pictureBox2.Image?.Dispose();this.pictureBox2.Image = img.Bitmap;}

5.也可以直接调用GrabCut方法，多迭代几次就可以，不过效率低

 

 private void button2_Click(object sender, EventArgs e){if(this.pictureBox1.Image==null){return;}float scaleX = (float)this.pictureBox1.Image.Width / pictureBox1.Width;float scaleY = (float)this.pictureBox1.Image.Height / pictureBox1.Height;Rectangle rect = new Rectangle((int)(Rect.Left * scaleX),(int)(Rect.Top * scaleY),(int)(Rect.Width * scaleX),(int)(Rect.Height * scaleY));var rectangle = new OpenCvSharp.Rect(rect.X, rect.Y, rect.Width, rect.Height);rect.Intersect(new Rectangle(0, 0, this.pictureBox1.Image.Width, this.pictureBox1.Image.Height));using Mat nimage = new Bitmap(this.pictureBox1.Image).ToMat();using Mat image = new Mat(nimage, rectangle);using Mat convertedImage = new Mat();Cv2.CvtColor(image, convertedImage, ColorConversionCodes.BGRA2BGR); // 将图像转换为CV_8UC3类型using Mat mask = new Mat(convertedImage.Size(), MatType.CV_8UC1, Scalar.Black);using Mat bgdModel = new Mat();using Mat fgdModel = new Mat();var r = new OpenCvSharp.Rect(0, 0, rectangle.Width-10, rectangle.Height-10);Cv2.GrabCut(convertedImage, mask, r, bgdModel, fgdModel, 20, GrabCutModes.InitWithRect);Cv2.Threshold(mask, mask, 2, 255, ThresholdTypes.Binary);using Mat foreground = new Mat();Cv2.BitwiseAnd(convertedImage, convertedImage, foreground, mask);// 创建一个具有 alpha 通道的新图像using Mat result = new Mat();Cv2.CvtColor(image, result, ColorConversionCodes.BGRA2BGR);Cv2.Rectangle(result, rectangle, new Scalar(0, 0, 0), 1); // 绘制矩形边界Cv2.Multiply(result, new Scalar(1, 1, 1, 0), result); // 将背景部分置为透明// 将前景部分复制到结果图像中image.CopyTo(result, mask);pictureBox2.Image?.Dispose();pictureBox2.Image = result.ToBitmap();image.Dispose();}

注：如果需要其它源码，请留言邮箱，我看到私发。