c# - C#木材缺陷检测算法的实现与测试
问题描述
我对此有一些疑问。
谁能告诉我“属于”I^2 的确切含义是什么？
我在 List 中有我的种子，但我不知道这是否正确。接下来，在第 2.3 点，我必须创建一个具有最小-最大点和隶属函数 bj 的超盒模糊集。
问题是当我在图片中有两个结时，我不能合理地将一个超级框分成例如两个超级框。在我看来，我应该有一些对我有帮助的课程。对于第 2.3 点，我只使用只有种子坐标的输入模式。
谁能告诉我我应该怎么做才能将一个超级盒分成两个超级盒？
谢谢。
namespace inz_test
{
class Input
{
Vector xy;
Color color;
double eukliedes;
public Vector XY
{
get { return xy; }
set { xy = value; }
}
public Color Pixel
{
get { return color; }
set { color = value;}
}
public double Eukliedes
{
get { return eukliedes; }
set { eukliedes = value; }
}
}
public class TestPhotos
{
public TestPhotos(string path, Bitmap img)
{
double counter = img.Height * img.Width;
double sumaR = 0, sumaG = 0, sumaB = 0;
int minR = 255, minG = 255, minB = 255;
//histogram array
var histogramR = new double[256];
var histogramG = new double[256];
var histogramB = new double[256];
var x_hArrayInputAll = new List<Vector>();
//test CSV files with data
StringBuilder csv = new StringBuilder();
StringBuilder csv1 = new StringBuilder();
/* ###################### */
//R,G,B value for
for (int y = 0; y < img.Height; y++)
{
for (int x = 0; x < img.Width; x++)
{
Color pixel = img.GetPixel(x, y);
//add R,G,B
histogramR[pixel.R]++;
histogramG[pixel.G]++;
histogramB[pixel.B]++;
//min for R,G and B channel
if (pixel.R < minR) minR = pixel.R;
if (pixel.G < minG) minG = pixel.G;
if (pixel.B < minB) minB = pixel.B;
//[0 1] range for all image pixel
Vector x_a = new Vector(Math.Round(((double)x / img.Width), 2),
Math.Round(((double)y / img.Height), 2));
x_hArrayInputAll.Add(x_a);
sumaR += pixel.R;
sumaG += pixel.G;
sumaB += pixel.B;
}
}
//csv path on .\Desktop\csv
string path1 = Environment.GetFolderPath(Environment.SpecialFolder.Desktop);
string filepath = path1 + "\\csv\\" + path + ".csv";
string filepath1 = path1 + "\\csv\\" + path + "_XY" + ".csv";
string filepath2 = path1 + "\\csv\\" + path + "1.jpg";
//cumulative histogram array
double[] cumulativeR = new double[256];
double[] cumulativeG = new double[256];
double[] cumulativeB = new double[256];
//cumulative[0]
cumulativeR[0] = histogramR[0] / counter;
cumulativeG[0] = histogramG[0] / counter;
cumulativeB[0] = histogramB[0] / counter;
//
for (var n = 1; n < histogramR.Length; n++)
{
cumulativeR[n] = (histogramR[n] / counter) + cumulativeR[n - 1];
cumulativeG[n] = (histogramG[n] / counter) + cumulativeG[n - 1];
cumulativeB[n] = (histogramB[n] / counter) + cumulativeB[n - 1];
}
double fiR, fiG, fiB;
double rhoR, rhoG, rhoB;
double lamR, lamG, lamB;
int sredniaR = (int)(sumaR/counter) ,sredniaG = (int)(sumaG / counter), sredniaB = (int)(sumaB / counter);
//set intensity level per channel
fiR = ((cumulativeR[(int)(sumaR / counter)]) * 0.01);
fiG = ((cumulativeG[(int)(sumaG / counter)]) * 0.01);
fiB = ((cumulativeB[(int)(sumaB / counter)]) * 0.01);
//set additional color intensity per channel
rhoR = ((fiR + (sredniaR)) / 2);
rhoG = ((fiG + (sredniaG)) / 2);
rhoB = ((fiB + (sredniaB)) / 2);
//lambda for channel
lamR = 50;
lamG = 50;
lamB = 50;
if (fiR < lamR) { fiR += rhoR; }
if (fiG < lamG) { fiG += rhoG; }
if (fiB < lamB) { fiB += rhoB; }
int maxH = 0, maxW = 0, minH = img.Height, minW = img.Width;
double euk = 0;
//Vector[] x_hArrayInput = new Vector[(int)counter];
var x_hArrayInput = new List<Vector>();
var inputList = new List<Input>();
//test
var list = new List<Tuple<Vector, Vector>>();
int c = 0;
double tau = 0.7;
double zFunction = 0;
Vector Vj1 = new Vector();
Vector Wj1 = new Vector();
for (int y = 0; y < img.Height; y++)
{
for (int x = 0; x < img.Width; x++)
{
Color pixel = img.GetPixel(x, y);
//if (pixel.R <= fiR)
//{
// if (pixel.G <= fiG)
// {
// if (pixel.B <= fiB)
// {
//
if ((pixel.R + pixel.G + pixel.B) / 3 <= (fiR + fiG + fiB) / 3)
{
Vector x_h = new Vector(Math.Round(((double)x / img.Width), 2), Math.Round(((double)y / img.Height), 2));
x_hArrayInput.Add(x_h);
Input i = new Input();
i.XY = x_h;
i.Pixel = pixel;
if (c>0)
{
var item = inputList[inputList.Count - 1];
var R = Math.Pow((item.Pixel.R - i.Pixel.R),2);
var G = Math.Pow((item.Pixel.G - i.Pixel.G),2);
var B = Math.Pow((item.Pixel.B - i.Pixel.B),2);
double distance = Math.Sqrt(R + G + B);
i.Eukliedes = distance;
double l = Math.Sqrt(195075);
if (distance == 0) { zFunction = 1; }
else if (0 <= distance && distance <= (l/2)) { zFunction = 1 - 2 * (Math.Pow((distance / l), 2)); }
else if ((l/2) <= distance && distance <= l) { zFunction = 2*(Math.Pow(((distance-l) / l), 2)); }
else if (distance == l) { zFunction = 0; }
csv.AppendLine(string.Format("{0};{1};{2};{3};{4};{5};{6}", i.XY.X, i.XY.Y, i.Pixel.R, i.Pixel.G, i.Pixel.B, distance, zFunction));
}
else
{
Vj1 = x_h;
Wj1 = x_h;
csv.AppendLine(string.Format("{0};{1};{2}", i.XY.X, i.XY.Y, i.Pixel));
}
c++;
inputList.Add(i);
//test red pixel on image
img.SetPixel(x, y, Color.Red);
//set max and min
if (maxH < y) { maxH = y; }
if (maxW < x) { maxW = x; }
if (minH > y) { minH = y; }
if (minW > x) { minW = x; }
}
}
}
double gamma = 1;
double bj;
Vector Vj = new Vector(Math.Round(((double)minW / img.Width), 2), Math.Round(((double)minH / img.Height), 2));
Vector Wj = new Vector(Math.Round(((double)maxW / img.Width), 2), Math.Round(((double)maxH / img.Height), 2));
double fun1, fun2, fun3, fun4;
Vector f_ramp_function_1 = new Vector();
Vector f_ramp_function_2 = new Vector();
csv1.AppendLine(string.Format("{0};{1};{2};{3};{4};{5};{6};{7};{8};{9};{10};{11};{12};{13};{14}", "x_hArrayInput[o].X", "x_hArrayInput[o].Y", "Vj.X", "Vj.Y", "Wj.X", "Wj.Y", "f_ramp_function_1.X", "f_ramp_function_1.Y", "f_ramp_function_2.X", "f_ramp_function_2.Y", "fun1", "fun2", "fun3", "fun4", "bj"));
for (int o = 0; o < x_hArrayInput.Count; o++)
{
f_ramp_function_1.X = x_hArrayInput[o].X - Wj.X;
f_ramp_function_1.Y = Vj.X - x_hArrayInput[o].X;
f_ramp_function_2.X = x_hArrayInput[o].Y - Wj.Y;
f_ramp_function_2.Y = Vj.Y - x_hArrayInput[o].Y;
//FOR 1-ST FUNCTION
if (f_ramp_function_1.X > 1)
{
fun1 = 1;
}
else if(0 <= f_ramp_function_1.X && f_ramp_function_1.X <= 1)
{
fun1 = f_ramp_function_1.X;
}
else
{
fun1 = 0;
}
if (f_ramp_function_1.Y > 1)
{
fun2 = 1;
}
else if (0 <= f_ramp_function_1.Y && f_ramp_function_1.Y <= 1)
{
fun2 = f_ramp_function_1.Y;
}
else
{
fun2 = 0;
}
//END 1-ST FUNCTION
//FOR 2-ND FUNCTION
if (f_ramp_function_2.X > 1)
{
fun3 = 1;
}
else if (0 <= f_ramp_function_2.X && f_ramp_function_2.X <= 1)
{
fun3 = f_ramp_function_2.X;
}
else
{
fun3 = 0;
}
if (f_ramp_function_2.Y > 1)
{
fun4 = 1;
}
else if (0 <= f_ramp_function_2.Y && f_ramp_function_2.Y <= 1)
{
fun4 = f_ramp_function_2.Y;
}
else
{
fun4 = 0;
}
//END 2-ND FUNCTION
bj = Math.Min((Math.Min((1-fun1),(1-fun2))),(Math.Min((1-fun3),(1-fun4))));
csv1.AppendLine(string.Format("{0};{1};{2};{3};{4};{5};{6};{7};{8};{9}", x_hArrayInput[o], Vj, Wj, f_ramp_function_1, f_ramp_function_2, fun1, fun2, fun3, fun4, bj));
}
//test hyperbox
for (int q = minW; q <= maxW; q++)
{
img.SetPixel(q, minH, Color.Black);
img.SetPixel(q, maxH, Color.Black);
//img.SetPixel(q, minH - 1, Color.Black);
}
for (int w = minH; w <= maxH; w++)
{
img.SetPixel(minW, w, Color.Black);
img.SetPixel(maxW, w, Color.Black);
//img.SetPixel(minW - 1, w, Color.Black);
}
File.AppendAllText(filepath1, csv1.ToString());
File.AppendAllText(filepath, csv.ToString());
img.Save(filepath2);
//System.GC.Collect();
}
}
}
解决方案
推荐阅读
- python - 如何在我的应用程序中包含 python IDE
- prolog - Prolog 剪切和失败未按预期工作
- javascript - 使用 Express 和 mongoDB 为用户提供高级访问权限并为其设置过期时间
- html - 为什么当我将鼠标悬停在这个具有 3d 旋转效果的元素上时，它会意外拉伸？
- django - 如何获取与产品类别页面上显示的产品关联的所有产品属性的列表？
- javascript - 角文本旋转器
- python - 如何拆分可选参数？
- pytorch - Pytorch 和 IOS 的演示不适用于 MacBook M1
- python - 为什么我的 matplotlib 脚本在 macbook 桌面上非常慢，但在外接显示器上运行良好？（Python、matplotlib、OSX）
- r - 有没有办法在服务器的过滤器语句中使用反应式单选按钮值