import numpy as np import cv2 import guidedFilter as gf import scipy.io as sio img = cv2.imread('./img_enhancement/tulips.bmp') / 255.0 test_img = img[:, :, 2] p = test_img r = 16 eps = 0.01 q = np.zeros(test_img.shape) # q,N,mean_I, mean_p, mean_Ip, cov_Ip,mean_II, var_I, a, b, mean_a, mean_b, q = gf.guidedfilter(test_img, p, r, eps); q = gf.guidedfilter(test_img, p, r, eps) I_enhanced = (test_img - q) * 5 + q # plt.subplot(1,3,1), plt.imshow(test_img) # # plt.subplot(132), plt.imshow(q) # # plt.subplot(133), plt.imshow(I_enhanced) # # plt.show() cv2.imshow('adsdfd', test_img) cv2.imshow('ada', q) cv2.imshow('sdfda', I_enhanced) cv2.waitKey(0) cv2.destroyAllWindows() # sio.savemat('saveddata.mat', {'N': N,'mean_I': mean_I, 'mean_p': mean_p, 'mean_Ip': mean_Ip, 'cov_Ip': cov_Ip, \
# example: detail enhancement # figure 6 in our paper import cv2 import numpy as np import guidedFilter as gf I = (cv2.imread( '/Users/bushanshan/Documents/Workspace/haze_removal/myCode/img_enhancement/tulips.bmp' )) / 255.0 p = I r = 16 eps = 0.01 q = np.zeros(I.shape) q[:, :, 0] = gf.guidedfilter(I[:, :, 2], p[:, :, 0], r, eps) q[:, :, 1] = gf.guidedfilter(I[:, :, 1], p[:, :, 1], r, eps) q[:, :, 2] = gf.guidedfilter(I[:, :, 0], p[:, :, 2], r, eps) I_enhanced = (I - q) * 5 + q print q.shape print I.shape print I_enhanced.shape cv2.imshow("I", I) cv2.imshow("q", q) cv2.imshow('I_enhanced', I_enhanced) cv2.waitKey(0) cv2.destroyAllWindows() # figure();
import cv2 import numpy as np import guidedFilter as gf I = cv2.imread("./cat.bmp") gray = cv2.cvtColor(I, cv2.COLOR_BGR2GRAY) # print gray.shape gray = I[:, :, 0] / 255.0 p = gray r = 4 # try r=2, 4, or 8 eps = 0.04 # % try eps=0.1^2, 0.2^2, 0.4^2 q = gf.guidedfilter(gray, p, r, eps) cv2.imshow("a", I) cv2.imshow("b", q) cv2.waitKey(0) cv2.destroyAllWindows()
import cv2 import guidedFilter as gf import scipy.io as sio img = cv2.imread('./img_enhancement/tulips.bmp') / 255.0; test_img = img[:, :, 2]; p = test_img; r = 16; eps = 0.01; q = np.zeros(test_img.shape); # q,N,mean_I, mean_p, mean_Ip, cov_Ip,mean_II, var_I, a, b, mean_a, mean_b, q = gf.guidedfilter(test_img, p, r, eps); q = gf.guidedfilter(test_img, p, r, eps); I_enhanced = (test_img - q) * 5 + q; # plt.subplot(1,3,1), plt.imshow(test_img) # # plt.subplot(132), plt.imshow(q) # # plt.subplot(133), plt.imshow(I_enhanced) # # plt.show() cv2.imshow('adsdfd', test_img) cv2.imshow('ada', q) cv2.imshow('sdfda', I_enhanced) cv2.waitKey(0) cv2.destroyAllWindows() # sio.savemat('saveddata.mat', {'N': N,'mean_I': mean_I, 'mean_p': mean_p, 'mean_Ip': mean_Ip, 'cov_Ip': cov_Ip, \
import cv2 import numpy as np import guidedFilter as gf I = cv2.imread('./cat.bmp') gray = cv2.cvtColor(I, cv2.COLOR_BGR2GRAY) # print gray.shape gray = I[:, :, 0] / 255.0 p = gray r = 4 # try r=2, 4, or 8 eps = 0.04 # % try eps=0.1^2, 0.2^2, 0.4^2 q = gf.guidedfilter(gray, p, r, eps) cv2.imshow('a', I) cv2.imshow('b', q) cv2.waitKey(0) cv2.destroyAllWindows()
# example: detail enhancement # figure 6 in our paper import cv2 import numpy as np import guidedFilter as gf I = (cv2.imread('/Users/bushanshan/Documents/Workspace/haze_removal/myCode/img_enhancement/tulips.bmp')) / 255.0; p = I; r = 16; eps = 0.01; q = np.zeros(I.shape); q[:, :, 0] = gf.guidedfilter(I[:, :, 2], p[:, :, 0], r, eps); q[:, :, 1] = gf.guidedfilter(I[:, :, 1], p[:, :, 1], r, eps); q[:, :, 2] = gf.guidedfilter(I[:, :, 0], p[:, :, 2], r, eps); I_enhanced = (I - q) * 5 + q; print q.shape print I.shape print I_enhanced.shape cv2.imshow("I", I) cv2.imshow("q", q) cv2.imshow('I_enhanced', I_enhanced) cv2.waitKey(0) cv2.destroyAllWindows() # figure();