基于Python实现遗传算法对图像进行去噪（源码+图片）.rar

# -*- coding:utf-8 -*- ''' 中值滤波器在椒盐噪声中的不同应用 ''' import numpy as np # Different applied median filter class Damf(object): def binary_matrix(self, A): '''Set the noise entry to 0.Set regular entry to 1''' B = np.zeros(np.shape(A)) B[(A == 255) | (A == 0)] = 0 B[(A != 255) & (A != 0)] = 1 return B def t_symmetric_pad_matrix(self, A, t): m, n = np.shape(A) img_pad = np.zeros((m + t, n + t)) img_pad[t:, t:] = A left_up = np.rot90(np.rot90(A[:t, :t])) up = np.flip(A[:t, :], 0) right_up = np.rot90(np.rot90(A[:t, -t:])) left = np.flip(A[:, 0:t], 1) right = np.flip(A[:, -t:], 1) left_down = np.rot90(np.rot90(A[-t:, :t])) dowm = np.flip(A[-t:, :], 0) right_down = np.rot90(np.rot90(A[-t:, -t:])) up_h = np.hstack((left_up, up, right_up)) # h：horizontal middle_h = np.hstack((left, A, right)) down_h = np.hstack((left_down, dowm, right_down)) img_pad = np.vstack((up_h, middle_h, down_h)) return img_pad def k_approximate_matrix(self, img_pad, i, j, k, t): k_approximate_mat = img_pad[(i + t - k):(i + t + k + 1), (j + t - k):(j + t + k + 1)] return k_approximate_mat def regular_row_matrix(self, img_k_approximate): H = img_k_approximate.flatten() H = np.sort(H) R = H[(H != 255) & (H != 0)] return R def process_image(self, nim): t1 = 3 img_DAMF = nim.copy() B = self.binary_matrix(nim) P_A_3 = self.t_symmetric_pad_matrix(nim, t1) P_B_3 = self.t_symmetric_pad_matrix(B, t1) m, n = np.shape(nim) for i in range(m): for j in range(n): if B[i, j] == 0: B_ij_1 = self.k_approximate_matrix(P_B_3, i, j, 1, t1) B_ij_2 = self.k_approximate_matrix(P_B_3, i, j, 2, t1) B_ij_3 = self.k_approximate_matrix(P_B_3, i, j, 3, t1) if B_ij_1.any() > 0: # 使用3x3板 A_ij_1 = self.k_approximate_matrix(P_A_3, i, j, 1, t1) R_ij_1 = self.regular_row_matrix(A_ij_1) mR_ij_1 = np.median(R_ij_1) img_DAMF[i, j] = mR_ij_1 elif B_ij_2.any() > 0: # 使用5x5模板 A_ij_2 = self.k_approximate_matrix(P_A_3, i, j, 2, t1) R_ij_2 = self.regular_row_matrix(A_ij_2) mR_ij_2 = np.median(R_ij_2) img_DAMF[i, j] = mR_ij_2 elif B_ij_3.any() > 0: # 使用7x7模板 A_ij_3 = self.k_approximate_matrix(P_A_3, i, j, 3, t1) R_ij_3 = self.regular_row_matrix(A_ij_3) mR_ij_3 = np.median(R_ij_3) img_DAMF[i, j] = mR_ij_3 B = self.binary_matrix(img_DAMF) t2 = 1 P_A_1 = self.t_symmetric_pad_matrix(img_DAMF, t2) P_B_1 = self.t_symmetric_pad_matrix(B, t2) for i in range(m): for j in range(n): if B[i, j] == 0: B_ij_1 = self.k_approximate_matrix(P_B_1, i, j, 1, t2) if B_ij_1.any() > 0: A_ij_1 = self.k_approximate_matrix(P_A_1, i, j, 1, t2) R_ij_1 = self.regular_row_matrix(A_ij_1) mR_ij_1 = np.median(R_ij_1) img_DAMF[i, j] = mR_ij_1 return img_DAMF