Figure 1.
Flowchart of the improved multiscale wavelet transform algorithm.
Figure 1.
Flowchart of the improved multiscale wavelet transform algorithm.
Figure 2.
The filtering effect of the simulated data of each algorithm when the variance of the Gaussian noise is 0.003, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2) and (g2) are the corresponding grayscale images.
Figure 2.
The filtering effect of the simulated data of each algorithm when the variance of the Gaussian noise is 0.003, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2) and (g2) are the corresponding grayscale images.
Figure 3.
The filtering effect of the simulated data of each algorithm when the variance of the Gaussian noise is 0.005, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2) and (g2) are the corresponding grayscale images.
Figure 3.
The filtering effect of the simulated data of each algorithm when the variance of the Gaussian noise is 0.005, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2) and (g2) are the corresponding grayscale images.
Figure 4.
The filtering effect of the simulated data of each algorithm when the variance of the Gaussian noise is 0.007, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2) and (g2) are the corresponding grayscale images.
Figure 4.
The filtering effect of the simulated data of each algorithm when the variance of the Gaussian noise is 0.007, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2) and (g2) are the corresponding grayscale images.
Figure 5.
The filtering effect of the simulated data of each algorithm when the variance of the Gaussian noise is 0.009, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2), and (g2) are the corresponding grayscale images.
Figure 5.
The filtering effect of the simulated data of each algorithm when the variance of the Gaussian noise is 0.009, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2), and (g2) are the corresponding grayscale images.
Figure 6.
Variation in image PSNR with noise after denoising with different algorithms.
Figure 6.
Variation in image PSNR with noise after denoising with different algorithms.
Figure 7.
Variation in image RMSE with noise after denoising by different algorithms.
Figure 7.
Variation in image RMSE with noise after denoising by different algorithms.
Figure 8.
Variation in image smoothness with noise after denoising by different algorithms.
Figure 8.
Variation in image smoothness with noise after denoising by different algorithms.
Figure 9.
Variation in image SSIM with noise after denoising by different algorithms.
Figure 9.
Variation in image SSIM with noise after denoising by different algorithms.
Figure 10.
The filtering effect of the measured data of each algorithm when the variance of the Gaussian noise is 0.05, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2) and (g2) are the corresponding grayscale images.
Figure 10.
The filtering effect of the measured data of each algorithm when the variance of the Gaussian noise is 0.05, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2) and (g2) are the corresponding grayscale images.
Figure 11.
The filtering effect of the measured data of each algorithm when the variance of the Gaussian noise is 0.08, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2) and (g2) are the corresponding grayscale images.
Figure 11.
The filtering effect of the measured data of each algorithm when the variance of the Gaussian noise is 0.08, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2) and (g2) are the corresponding grayscale images.
Figure 12.
The filtering effect of the measured data of each algorithm when the variance of the Gaussian noise is 0.11, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2), and (g2) are the corresponding grayscale images.
Figure 12.
The filtering effect of the measured data of each algorithm when the variance of the Gaussian noise is 0.11, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2), and (g2) are the corresponding grayscale images.
Figure 13.
The filtering effect of the measured data of each algorithm when the variance of the Gaussian noise is 0.14, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2), and (g2) are the corresponding grayscale images.
Figure 13.
The filtering effect of the measured data of each algorithm when the variance of the Gaussian noise is 0.14, (a1) represents the original magnetic anomaly benchmark image, (b1) the added noise magnetic anomaly benchmark image, (c1) the Gaussian filter denoising effect image, (d1) the mean filtering effect image, (e1) the soft-thresholding wavelet filtering effect image, (f1) the hard-thresholding wavelet filtering effect image, (g1) the filtering effect image of this paper’s algorithm, respectively, and (a2), (b2), (c2), (d2), (e2), (f2), and (g2) are the corresponding grayscale images.
Figure 14.
Variation in image PSNR with noise after denoising with different algorithms.
Figure 14.
Variation in image PSNR with noise after denoising with different algorithms.
Figure 15.
Variation in image RMSE with noise after denoising by different algorithms.
Figure 15.
Variation in image RMSE with noise after denoising by different algorithms.
Figure 16.
Variation in image smoothness with noise after denoising by different algorithms.
Figure 16.
Variation in image smoothness with noise after denoising by different algorithms.
Figure 17.
Variation in image SSIM with noise after denoising by different algorithms.
Figure 17.
Variation in image SSIM with noise after denoising by different algorithms.