新手学习opencv七：svm

灰太狼 2022-08-20 12:13 184阅读 0赞

新手学习opencv七：svm

新手学习opencv，svm。

> **1）opencv中svm参数和函数**
> 
> **svm参数：**
> 
> CvSVMParams::CvSVMParams() :
> 
>     svm\_type(CvSVM::C\_SVC), kernel\_type(CvSVM::RBF), degree(0),
> 
>     gamma(1), coef0(0), C(1), nu(0), p(0), class\_weights(0)
> 
> \{
> 
>     term\_crit = cvTermCriteria( CV\_TERMCRIT\_ITER+CV\_TERMCRIT\_EPS, 1000, FLT\_EPSILON );
> 
> \}
> 
>  
> 
> svm\_type表示SVM类型：
> 
> \{C\_SVC=100, NU\_SVC=101, ONE\_CLASS=102, EPS\_SVR=103, NU\_SVR=104 \};//SVC是SVM分类器，SVR是SVM回归
> 
> kernel\_type表示核函数的类型：
> 
>  \{ LINEAR=0, POLY=1, RBF=2, SIGMOID=3 \}; //提供四种核函数，分别是线性，多项式，径向基，sigmoid型函数。
> 
> degree针对多项式核函数degree的设置，
> 
> gamma针对多项式/rbf/sigmoid核函数的设置
> 
> oef0针对多项式/sigmoid核函数的设置
> 
> 默认值degree = 0，gamma = 1，coef0 = 0
> 
> **svm训练函数**
> 
> bool CvSVM::train( const CvMat\* \_train\_data, const CvMat\* \_responses,
> 
> const CvMat\* \_var\_idx, const CvMat\* \_sample\_idx, CvSVMParams \_params )
> 
> 用的时候：svm.train(dataMat, labelMat, NULL, NULL, param);
> 
> **svm预测函数**
> 
> svm.predict(testMat);
> 
> 支持向量机的个数
> 
> svm.get\_support\_vector\_count();
> 
> 支持向量机
> 
> svm.get\_support\_vector(i)
> 
> 保存训练结果
> 
> svm.save("svmResult.txt");
> 
> **2）用法流程：**
> 
> 1.设置svm参数
> 
> 2.输入训练样本矩阵
> 
> 3.训练样本
> 
> 4.输入测试数据，根据训练结果分类
> 
>     /svm/
>     #include "cv.h"
>     #include <highgui.h>
>     #include "ml.h"
>     #include "iostream"
>     using namespace std;
>     double inputArr[10][13] = 
>     {
>     	1,0.708333,1,1,-0.320755,-0.105023,-1,1,-0.419847,-1,-0.225806,0,1, 
>     	-1,0.583333,-1,0.333333,-0.603774,1,-1,1,0.358779,-1,-0.483871,0,-1,
>     	1,0.166667,1,-0.333333,-0.433962,-0.383562,-1,-1,0.0687023,-1,-0.903226,-1,-1,
>     	-1,0.458333,1,1,-0.358491,-0.374429,-1,-1,-0.480916,1,-0.935484,0,-0.333333,
>     	-1,0.875,-1,-0.333333,-0.509434,-0.347032,-1,1,-0.236641,1,-0.935484,-1,-0.333333,
>     	-1,0.5,1,1,-0.509434,-0.767123,-1,-1,0.0534351,-1,-0.870968,-1,-1,
>     	1,0.125,1,0.333333,-0.320755,-0.406393,1,1,0.0839695,1,-0.806452,0,-0.333333,
>     	1,0.25,1,1,-0.698113,-0.484018,-1,1,0.0839695,1,-0.612903,0,-0.333333,
>     	1,0.291667,1,1,-0.132075,-0.237443,-1,1,0.51145,-1,-0.612903,0,0.333333,
>     	1,0.416667,-1,1,0.0566038,0.283105,-1,1,0.267176,-1,0.290323,0,1
>     };
>     double testArr[]=
>     {
>     	//0.25,1,1,-0.226415,-0.506849,-1,-1,0.374046,-1,-0.83871,0,-1
>     	1.29,1,1,-0.132075,-0.237443,-1,1,0.51145,-1,-0.612903,0,0.333333
>     };
>     int main()
>     {
>     	//1.svm参数设置
>     	CvSVM svm;
>     	CvSVMParams param;
>     	param.svm_type = 100;
>     	param.kernel_type = 1;
>     	param.degree = 2;
>     	param.gamma = 1;
>     	param.coef0 = 0;
>     	//2.输入样本矩阵
>     	CvMat *trainMat = cvCreateMat(10, 12, CV_32FC1);
>     	CvMat *labelMat = cvCreateMat(10, 1, CV_32SC1);//标签要有符号，代表正负
>     	for (int i=0; i<10; i++)
>     	{
>     		for (int j=0; j<12; j++)
>     		{
>     			cvSetReal2D(trainMat, i, j, inputArr[i][j+1]);
>     		}
>     		cvSetReal2D(labelMat, i, 0, inputArr[i][0]);
>     	}
>     	//3训练样本
>     	//计算时间
>     	double t = (double)cvGetTickCount();
>     	svm.train(trainMat, labelMat, NULL, NULL, param);
>     	t = (double)cvGetTickCount() - t;
>     	double timecost = t/(cvGetTickFrequency()*1000); 
>     	cout<<"svm train耗时："<<timecost<<"ms"<<endl;
>     	//保存训练结果
>     	svm.save("svmResult.txt");
>     	CvMat *testMat = cvCreateMat(1, 12, CV_32FC1);
>     	for (int i=0; i<12; i++)
>     	{
>     		cvSetReal2D(testMat, 0, i, testArr[i]);
>     	}
>     
>     	//4预测分类
>     	float flag = 0;
>     	flag = svm.predict(testMat);
>     	int c = svm.get_support_vector_count();
>     	cout<<"svm个数"<<c<<endl;
>     	for (int i=0; i<c; i++)
>     	{
>     		const float* v = svm.get_support_vector(i);
>     		for(int i=0;i<12;i++)
>     		{
>     			cout<<*(v+i)<<",";
>     		}
>     		cout<<endl;
>     	}
>     	
>     	cout<<"testMat分类结果："<<flag<<endl;
>     	system("pause");
>     	cvReleaseMat(&trainMat);
>     	cvReleaseMat(&labelMat);
>     	cvReleaseMat(&testMat);
>     	return 0;
>     
>     }
> 
>   
>   
>

![Center][]

[Center]: /images/20220731/f86fda52fe914c9d9e5b97d2f84091d8.png