opencv中自适应阈值函数的实现（c++）

根据《面向飞机蒙皮接缝的线结构光检测技术研究_张卡》论文中的原理，编写了自适应阈值函数

原理：

 //计算灰度最大最小值
 void MaxGrayValue(Mat image,int &max,int &min)
 {
     max = ;
     min = ;
     Mat *im = reinterpret_cast<Mat*>((void*)&image);
     int height = image.rows;
     int width = image.cols;
     int N = height * width;
     //cout << "125" << " " << "125" << " " << static_cast<int>(im->at<uchar>(Point(182, 190))) << endl;
     for (int i = ; i < height; i++)
     {
         for (int j = ; j < width; j++)
         {
             if ((static_cast<int>(im->at<uchar>(Point(j, i))))>max)
                 max = static_cast<int>(im->at<uchar>(Point(j, i)));
             else if ((static_cast<int>(im->at<uchar>(Point(j, i))))<min)
                 min = static_cast<int>(im->at<uchar>(Point(j, i)));
         }
     }
 }

 //自适应阈值 注：传入的只能是灰度图像
 void adapthreshold(Mat &image)
 {
     int max, min;    //保存最大和最小灰度值
     MaxGrayValue(image, max, min);   //计算出最大最小灰度值
     int grayvalue;    //用于暂时存取每个点的灰度值
     int Threshold = (max + min)/;  //初始阈值
     int Temp_Threshold = Threshold; //用于存取每次迭代后的阈值
     int big_than_th;       //用于累加大于阈值的灰度值
     int bigcount;           //用于存放大于阈值灰度值的像素点数量
     int sma_than_th;       //用于累加小于阈值的灰度值
     int smacount;           //用于存放小于阈值灰度值的像素点数量
     Mat *im = reinterpret_cast<Mat*>((void*)&image);    //获取像素点信息
     //Mat *im = &image;
     int height = image.rows;   //获取图像高度
     int width = image.cols;       //获取图像宽度

     for (int k = ; k < ; k++)   //计算阈值，最大迭代50次
     {
         big_than_th = ;
         bigcount = ;
         sma_than_th = ;
         smacount = ;
         for (int i = ; i < height; i++)
         {
             for (int j = ; j < width; j++)
             {
                 grayvalue = static_cast<int>(im->at<uchar>(Point(j, i)));  //获取指定点灰度值
                 if (grayvalue>Threshold)   //如果灰度值大于阈值，加到总灰度值，数量+1
                 {
                     big_than_th += grayvalue;
                     bigcount++;
                 }
                 else if (grayvalue < Threshold) //如果灰度值小于阈值，加到总灰度值，数量+1
                 {
                     sma_than_th += grayvalue;
                     smacount++;
                 }
             }
         }
         Temp_Threshold = ((big_than_th*1.0 / bigcount) + (sma_than_th*1.0 / smacount)) / ;  //计算新的阈值
         if (abs(Temp_Threshold - Threshold) < )   //如果新的阈值和旧的阈值差值的绝对值小于10，则退出
             break;
         else
             Threshold = Temp_Threshold;    //否则旧阈值等于新阈值
     }
     for (int i = ; i < height; i++)     //对图像进行二值化处理
     {
         uchar* data = image.ptr<uchar>(i);
         for (int j = ; j < width; j++)
         {
             grayvalue = static_cast<int>(im->at<uchar>(Point(j, i)));
             if (grayvalue>Threshold)
                 data[j] = ;
             else if (grayvalue < Threshold)
                 data[j] = ;
         }
     }
 }