基于C++和QT开发的刀具破损状态检测软件源码（SVM进行判别）.zip

#include "RT_TCM_AE.h" //para1：信号 para2：平滑数 RT_TCM_AE::RT_TCM_AE(vector<float>Ae_data, int filter_size, int Frequency, int cut_threshold) :PI(3.1416), Frequency_length(65536) { this->len = filter_size; this->_Frequency = Frequency; this->_cut_threshold = cut_threshold; filter_res.clear(); raw_data.clear(); time_value.clear(); WPT_res.clear(); raw_data = Ae_data; filter_sum = 0; //设置频域及时频域分析时所需的数据长度,其必须是2的幂级数 //数据过长会导致计算频域、时频域特征时堆栈溢出 if (_Frequency < Frequency_length) { for (int i = 15; i > 0; --i) { if (pow(2, i) < _Frequency) { Frequency_length = pow(2, i); break; } } } } ///功能描述：无参构造函数：调用SVM类时初始化 ///param [I] 无 ///返回值： 无 ///备注：无 RT_TCM_AE::RT_TCM_AE() :PI(3.1416) { } //均值滤波+峰值计算 void RT_TCM_AE::cal_Peak() { int length = raw_data.size(); for (int i = 0; i < length; ++i) { filter_res.push_back(this->next(raw_data[i])); } ////峰值提取 vector<float>peak_points = calculate_Peak(filter_res); // cout << "峰值点数为：" << peak_points.size() << " "; Fluctuation = calcalate_FD(peak_points, 0.7, 1.3); } //均值滤波步骤 float RT_TCM_AE::next(float val) { que.push(val); filter_sum += val; while(que.size() > len) { filter_sum -= que.front();//超出窗口容量，先把队首元素获取，在sum里减去 que.pop();//再把队首元素弹出 } return filter_sum / que.size(); } //获取峰值波动特征 float RT_TCM_AE::get_FD() { return Fluctuation; } vector<float> RT_TCM_AE::calculate_Peak(vector<float> smooth_data) { int length = smooth_data.size(); vector<float>peak_ans; float avg = -1; float sum = 0; for (int i = 0; i < length; ++i) { sum += smooth_data[i]; } avg = float(sum / length); vector<int> coordinate_x; for (int i = 1; i < length - 1; ++i) { if (smooth_data[i] > smooth_data[i - 1] && smooth_data[i] > smooth_data[i + 1] && smooth_data[i] >= avg) {// if (!coordinate_x.empty()&&(i - coordinate_x.back() < 2000)) { if (smooth_data[i] > peak_ans.back()) { peak_ans.pop_back(); coordinate_x.pop_back(); coordinate_x.push_back(i); peak_ans.push_back(smooth_data[i]); } else { continue; } } else { coordinate_x.push_back(i); peak_ans.push_back(smooth_data[i]); } } } return peak_ans; } float RT_TCM_AE::calcalate_FD(vector<float> data, float low_hold, float high_hold) { int brow = data.size(); int vebral = 0; float val = 0; for (int i = 0; i < brow - 1; ++i) { val = data[i] / data[i + 1]; if (val<=low_hold || val >= high_hold) { vebral+=1; } } //cout<<"超出范围为"<< vebral<<" "; float ans = float(vebral)/float(brow); return ans; } //判断是否加工区域 bool RT_TCM_AE::stage_cutting() { float process_value = 0; float decision_threshold = 0.15;//设定加工下限阈值 int datasize = raw_data.size(); for (int j = 0; j < datasize; ++j) { process_value += raw_data[j]; } //计算均值 float average = process_value / (float)datasize; return average > decision_threshold; } //获取时域特征 vector<float> RT_TCM_AE::getTime() { return time_value; } //计算时域特征 void RT_TCM_AE::cal_Time() { float process_value = 0; float Rms_value = 0; float max_value = -1; int datasize = raw_data.size(); for (int j = 0; j < datasize; ++j) { if (raw_data[j] > max_value)max_value = raw_data[j]; process_value += raw_data[j]; Rms_value += raw_data[j] * raw_data[j]; } //计算均值 float average = process_value / (float)datasize; //计算均方根 float rms = Rms_value / (float)datasize; //计算峰值因子 float perk = max_value / rms; time_value.push_back(average); time_value.push_back(rms); time_value.push_back(perk); } ////频域分析 inline void RT_TCM_AE::swap(float &a, float &b) { float t; t = a; a = b; b = t; } void RT_TCM_AE::bitrp(float xreal[], float ximag[], int n) { // 位反转置换 Bit-reversal Permutation int i, j, a, b, p; for (i = 1, p = 0; i < n; i *= 2) { p++; } for (i = 0; i < n; i++) { a = i; b = 0; for (j = 0; j < p; j++) { b = (b << 1) + (a & 1); // b = b * 2 + a % 2; a >>= 1; // a = a / 2; } if (b > i) { swap(xreal[i], xreal[b]); swap(ximag[i], ximag[b]); } } } void RT_TCM_AE::FFT(float xreal[], float ximag[], int n) { // 快速傅立叶变换，将复数 x 变换后仍保存在 x 中，xreal, ximag 分别是 x 的实部和虚部 float *wreal = new float[_Frequency / 2]; float *wimag = new float[_Frequency / 2]; //float wreal[frequency / 2], wimag[frequency / 2]; float treal, timag, ureal, uimag, arg; int m, k, j, t, index1, index2; bitrp(xreal, ximag, n); // 计算 1 的前 n / 2 个 n 次方根的共轭复数 W'j = wreal [j] + i * wimag [j] , j = 0, 1, ... , n / 2 - 1 arg = -2 * PI / n; treal = cos(arg); timag = sin(arg); wreal[0] = 1.0; wimag[0] = 0.0; for (j = 1; j < n / 2; j++) { wreal[j] = wreal[j - 1] * treal - wimag[j - 1] * timag; wimag[j] = wreal[j - 1] * timag + wimag[j - 1] * treal; } for (m = 2; m <= n; m *= 2) { for (k = 0; k < n; k += m) { for (j = 0; j < m / 2; j++) { index1 = k + j; index2 = index1 + m / 2; t = n * j / m; // 旋转因子 w 的实部在 wreal [] 中的下标为 t treal = wreal[t] * xreal[index2] - wimag[t] * ximag[index2]; timag = wreal[t] * ximag[index2] + wimag[t] * xreal[index2]; ureal = xreal[index1]; uimag = ximag[index1]; xreal[index1] = ureal + treal; ximag[index1] = uimag + timag; xreal[index2] = ureal - treal; ximag[index2] = uimag - timag; } } } //清除一维数组内存 delete[] wreal; wreal = nullptr; delete[] wimag; wimag = nullptr; } void RT_TCM_AE::cal_FFT() { //动态数据创建实数、复数 float *xreal = new float[_Frequency]; float *ximag = new float[_Frequency]; // //float xreal[frequency] = {}, ximag[frequency] = {}; //float xreal[frequency] = {}, ximag[frequency] = {0}; int n, i= Frequency_length; for (int j = 0; j < i; ++j) { xreal[j]=raw_data[j]; ximag[j]=0; } n = i; // 要求 n 为 2 的整数幂 while (i > 1) { if (i % 2) { printf( "%d is not a power of 2! ", n); exit(1); } i /= 2; } FFT(xreal, ximag, n); //printf( "FFT: i real imag\n"); float sum = 0; for (i = 0; i < n; i++) { sum += xreal[i]; //printf( "%4d %8.4f %8.4f\n", i, xreal[i], ximag[i]); } fft_energy = sum / (float)n; //清除一维数组内存 delete[] xreal; xreal = nullptr; delete[] ximag; ximag = nullptr; } float RT_TCM_AE::get_FFT() { return fft_energy; } //计算小波包能量比结果 void RT_TCM_AE::cal_WPT() { vector<double>signal; int data_size = Frequency_length; for (int i = 0;i < data_size; ++i) { signal.push_back(raw_data[i]); } std::vector<double> reconstructed_signal; reconstructed_signal.resize(signal.size()); // Load well-known wavelet coefficients. Wavelet wavelet; Wavelet::GetWaveletCoefficients(&wavelet, Wavelet::WaveletType::Daubechies, 5); // Construct wavelet packet tree of height = 3. // This tree has 4 wavelet levels, and 4 leaves. WaveletPacketTree tree(4, &wavelet); tree.SetRootSignal(&signal); tree.Decompose(); // Reconstruct signal one wavelet level at a time. //cout << "小波包叶子为：" << tree.GetWaveletLevelCount() << endl; vector<double>energy; for (size_t i = 0; i < tree.GetWaveletLevelCount(); i++) { tree.Reconstruct(i); auto root_signal = tree.GetRootSignal(); // Add the coefficients of wavelet level i to the reconstructed signal. std::transform(reconstructed_signal.cbegin(), reconstructed_signal.cend(), root_signal->cbegin(), reconstructed_signal.begin(), std::plus<double>()); //root_signal:最后一层小波包重构信号 double sum = 0; for_each(root_signal->cbegin(), root_signal->cend(), [&](double x) { sum += x; }); sum = sqrtf(sum); energy.push_back(sum); } //for_each(energy.begin(), energy.end(), [](float x) { // cout << x << " "; //}); double all_sum = 0; for_each(energy.begin(), energy.end(), [&](double x) { all_sum += x; }); WPT_res.push_back(energy[2]/all_sum*100); WPT_res.push_back(energy[4] / all