Founded in 1992，
bimonthly
sponsor：
Institute of Geophysics, China Earthquake Administration
Nuctech Company Limited
Nuctech Company Limited
NoticeMore>
linksMore>
Application of Joint Inversion of Different Electrode Arrays in Ancient Mausoleum Detection
Electrical resistivity tomography is a popular geophysical method and has been applied in shallow exploration, involving hydrology, archaeology, and geology, in recent years. To enhance the resolution of electrical resistivity tomography and deal with complex geological settings, we propose the weighted combined inversion of different electrode arrays based on the Jacobian matrix, and then, taking Wenner and dipole-dipole datasets as examples, test its effectiveness on synthetic models and a field case of detecting ancient mausoleum. The results show that the resolution of the weighted combined inversion results is superior to that of a single electrode array in transverse and longitudinal directions, and in the field case, it is demonstrated that the weighted combined inversion algorithm can alleviate the inherent defects of U-shaped electrode array, reduce the ambiguity of inversion, and better constrain the width of the mausoleum.
Coded Aperture Computed Tomography Via Generative Adversarial U-net
Generative adversarial U-net for coded aperture computed tomography (CT) is proposed in this paper to alleviate the tradeoff between the non-continuous sparse projections and the ill-posedness iterative reconstruction problem. A non-continuous sparse projection model is presented based on generative adversarial U-net and the corresponding joint penalty function is formulated. Simulations using real datasets show that CT images with 256×256 pixels can be reconstructed with peak signal-to-noise ration more than 30 dB at only 5% transmittance. Furthermore, the computational time in the reconstructions is reduced by two orders of magnitude when compared with the state-of-the-art iterative algorithms in coded aperture computed tomography.
Sparse CT Reconstruction Based on Adversarial Residual Dense Deep Neural Network
To solve the problem of severe streak artifacts in sparse-view computed tomography (CT) reconstruction, in this paper we propose a method which is based on the adversarial residual dense deep neural network to acquire high-quality sparse-view CT reconstruction. The UNet that combines residual connectioin, dense connection, adversarial mechanism and attention mechanism is designed, which is trained through large-scale training data composed of streak artifact images and high-quality images to suppress streak artifacts. First, the filtered back projection (FBP) algorithm is used to reconstruct CT images with streak artifacts from sparse projections, then these images are inputed into the deep network, which can suppress streak artifacts to output high-quality images. The experimental results show that, compared with the existing deep learning algorithms, the image reconstructed by the proposed new network possesses higher accuracy and can suppress streak artifacts better.
Research on Material Decomposition of Dual-energy CT Image Based on Iterative Residual Network
Dual energy computed tomography (DECT) plays an important role in the application of advanced imaging due to its material decomposition capability. Image domain decomposition can directly invert CT images through by linear matrix, but the decomposed material images will be seriously affected by noise and artifacts. Although various regularization methods have been proposed to solve this problem, they still face two challenges: tedious parameter adjustment and the loss of image details resulted from over-smoothing. Therefore, in this paper we proposes a dual energy CT image material decomposition algorithm based on iterative residual network. Direct inversion is used as the initial base image, and a stacking two-channel convolutional neural network is used to replace the regularization items in the iterative decomposition model to form a deep iterative decomposition network. This method can realize material decomposition and noise suppression simultaneously. Experimental results show that the iterative residual network proposed in this paper is superior to other comparison methods and can effectively suppress noise and artifacts while maintaining the edge details of the base image.
Advances in Research on Low-dose CT Imaging Algorithm Based on Deep Learning
Computed tomography (CT) is widely used in clinical diagnosis because of its fast imaging speed and high resolution. However, higher doses of radiation will cause damages to human tissues and organs, while lower doses will lead to serious deterioration of imaging quality. In order to solve the above contradiction, researchers have focused on the low-dose CT imaging technology to study how to reduce the harm caused by radiation to the human body to the greatest extent under the condition of ensuring the imaging quality to meet the needs of clinical diagnosis. In recent years, deep learning has developed rapidly in the field of artificial intelligence, and has been widely used in image processing, pattern recognition, signal processing fields. Driven by big data, LDCT imaging algorithms based on deep learning have made great progress. This paper studies the development of low-dose CT imaging algorithms in recent years in terms of three aspects: the process of CT imaging, the noise modeling of low-dose CT, and the design of imaging algorithms. In particular, the imaging algorithms in the field of deep learning are systematically elaborated and analyzed. Finally, future developments in the field of LDCT image artifact suppression are also prospected.
Articles Online First have been peer-reviewed and accepted, which are not yet assigned to volumes /issues, but are citable by Digital Object Identifier (DOI).
Display Method:
Display Method:
2022, 31(3): 269-279.
doi: 10.15953/j.ctta.2021.075
Abstract:
In the process of numerical simulation, boundary reflection is an important factor which affect the numerical simulation results. The actual underground medium holds anisotropic characteristics. The traditional perfectly matched layer boundary (PML) shows good effect on small incident angle seismic waves, yet it can not effectively absorb low-frequency waves and large angle incident waves. To solve the problem of boundary reflection, in this paper, we propose a combined boundary condition using convolution perfectly matched layer (CPML) and eigenvalue analysis method to be applied in the numerical simulation of finite difference method in frequency space domain. The numerical simulation experiment and boundary reflection absorption effect analysis of the combined boundary condition verify that the proposed method is a reliable artificial absorption boundary condition, which can effectively suppress the boundary reflection generated in the process of wave field simulation.
In the process of numerical simulation, boundary reflection is an important factor which affect the numerical simulation results. The actual underground medium holds anisotropic characteristics. The traditional perfectly matched layer boundary (PML) shows good effect on small incident angle seismic waves, yet it can not effectively absorb low-frequency waves and large angle incident waves. To solve the problem of boundary reflection, in this paper, we propose a combined boundary condition using convolution perfectly matched layer (CPML) and eigenvalue analysis method to be applied in the numerical simulation of finite difference method in frequency space domain. The numerical simulation experiment and boundary reflection absorption effect analysis of the combined boundary condition verify that the proposed method is a reliable artificial absorption boundary condition, which can effectively suppress the boundary reflection generated in the process of wave field simulation.
2022, 31(3): 280-292.
doi: 10.15953/j.ctta.2022.008
Abstract:
Electrical resistivity tomography is a popular geophysical method and has been applied in shallow exploration, involving hydrology, archaeology, and geology, in recent years. To enhance the resolution of electrical resistivity tomography and deal with complex geological settings, we propose the weighted combined inversion of different electrode arrays based on the Jacobian matrix, and then, taking Wenner and dipole-dipole datasets as examples, test its effectiveness on synthetic models and a field case of detecting ancient mausoleum. The results show that the resolution of the weighted combined inversion results is superior to that of a single electrode array in transverse and longitudinal directions, and in the field case, it is demonstrated that the weighted combined inversion algorithm can alleviate the inherent defects of U-shaped electrode array, reduce the ambiguity of inversion, and better constrain the width of the mausoleum.
Electrical resistivity tomography is a popular geophysical method and has been applied in shallow exploration, involving hydrology, archaeology, and geology, in recent years. To enhance the resolution of electrical resistivity tomography and deal with complex geological settings, we propose the weighted combined inversion of different electrode arrays based on the Jacobian matrix, and then, taking Wenner and dipole-dipole datasets as examples, test its effectiveness on synthetic models and a field case of detecting ancient mausoleum. The results show that the resolution of the weighted combined inversion results is superior to that of a single electrode array in transverse and longitudinal directions, and in the field case, it is demonstrated that the weighted combined inversion algorithm can alleviate the inherent defects of U-shaped electrode array, reduce the ambiguity of inversion, and better constrain the width of the mausoleum.
2022, 31(3): 293-304.
doi: 10.15953/j.ctta.2021.017
Abstract:
Due to that the main reservoirs of the X gas field in the East China Sea were deeply buried and have large lateral changes, the conventional seismic data resulted in poor quality and low resolution, which couldn’t meet the increasingly refined geological requirements in exploration and development. Seismic data with wideband and wide azimuth was obtained by using the acquisition method of three ships and four sources with oblique cables, which held the characteristics of high resolution, high signal-to-noise ratio and high fidelity. By taking advantage of the superior information of wideband and wide azimuth seismic data which was high-resolutional and anisotropic, combined with the simultaneous prestack inversion, the inversion body of sensitive elastic parameters of channel sand bodies in different azimuths can be obtained, and superimpose multiple azimuth inversion bodies perpendicular to the direction of the river channel to carry out fine predictions of channel sand bodies. Compared the conventional seismic data, reservoir inversion based on wideband and wide azimuth seismic data improved the prediction accuracy of channel sand bodies, laying a foundation for the progressive exploration and development of the X gas field in the East China Sea.
Due to that the main reservoirs of the X gas field in the East China Sea were deeply buried and have large lateral changes, the conventional seismic data resulted in poor quality and low resolution, which couldn’t meet the increasingly refined geological requirements in exploration and development. Seismic data with wideband and wide azimuth was obtained by using the acquisition method of three ships and four sources with oblique cables, which held the characteristics of high resolution, high signal-to-noise ratio and high fidelity. By taking advantage of the superior information of wideband and wide azimuth seismic data which was high-resolutional and anisotropic, combined with the simultaneous prestack inversion, the inversion body of sensitive elastic parameters of channel sand bodies in different azimuths can be obtained, and superimpose multiple azimuth inversion bodies perpendicular to the direction of the river channel to carry out fine predictions of channel sand bodies. Compared the conventional seismic data, reservoir inversion based on wideband and wide azimuth seismic data improved the prediction accuracy of channel sand bodies, laying a foundation for the progressive exploration and development of the X gas field in the East China Sea.
2022, 31(3): 305-316.
doi: 10.15953/j.ctta.2021.088
Abstract:
The main target layer of the N structure in the East China Sea is a delta subaqueous distributary channel sand body developed under a strong hydrodynamic environment. The distribution of planar sand layer is discontinuous and the lateral heterogeneity is very strong. Under the influence of deep burial compaction and diagenesis, the reservoir is characterized by low porosity and low permeability, and the properties of rock-physics are overlapped seriously. In addition, the lack of large angle information of deep seismic data is a common problem. It is of great importance to implement the fluid distribution range of tight reservoir in the study area for the design and deployment of exploration and development. In this paper, a new method of deep seismic fluid description is introduced based on direct inversion of lame parameters. Through qualitative and quantitative analysis of rock-physics of measured well data, optimal highly sensitive hydrocarbon detection factor is selected. Furthermore, the AVO properties of Lambda parameters are extracted from the pre-stack trace set by combining with the parametric equations of the two AVO models of Lambda parameters. Then, the AVO properties are directly transformed into the interlayer elastic information by using the colored inversion technique. Finally, the seismic fluid sensitive elastic data is obtained to guide the seismic fluid description. The practical application shows that the hydrocarbon detection results of this method are compatible with the logging interpretation achievement, and can effectively describe the low-porosity and low-permeability reservoirs fluid development law of the study area, and can provide important technical support for the discovery of oil and gas resources in new fields.
The main target layer of the N structure in the East China Sea is a delta subaqueous distributary channel sand body developed under a strong hydrodynamic environment. The distribution of planar sand layer is discontinuous and the lateral heterogeneity is very strong. Under the influence of deep burial compaction and diagenesis, the reservoir is characterized by low porosity and low permeability, and the properties of rock-physics are overlapped seriously. In addition, the lack of large angle information of deep seismic data is a common problem. It is of great importance to implement the fluid distribution range of tight reservoir in the study area for the design and deployment of exploration and development. In this paper, a new method of deep seismic fluid description is introduced based on direct inversion of lame parameters. Through qualitative and quantitative analysis of rock-physics of measured well data, optimal highly sensitive hydrocarbon detection factor is selected. Furthermore, the AVO properties of Lambda parameters are extracted from the pre-stack trace set by combining with the parametric equations of the two AVO models of Lambda parameters. Then, the AVO properties are directly transformed into the interlayer elastic information by using the colored inversion technique. Finally, the seismic fluid sensitive elastic data is obtained to guide the seismic fluid description. The practical application shows that the hydrocarbon detection results of this method are compatible with the logging interpretation achievement, and can effectively describe the low-porosity and low-permeability reservoirs fluid development law of the study area, and can provide important technical support for the discovery of oil and gas resources in new fields.
2022, 31(3): 317-327.
doi: 10.15953/j.ctta.2021.070
Abstract:
Generative adversarial U-net for coded aperture computed tomography (CT) is proposed in this paper to alleviate the tradeoff between the non-continuous sparse projections and the ill-posedness iterative reconstruction problem. A non-continuous sparse projection model is presented based on generative adversarial U-net and the corresponding joint penalty function is formulated. Simulations using real datasets show that CT images with 256×256 pixels can be reconstructed with peak signal-to-noise ration more than 30 dB at only 5% transmittance. Furthermore, the computational time in the reconstructions is reduced by two orders of magnitude when compared with the state-of-the-art iterative algorithms in coded aperture computed tomography.
Generative adversarial U-net for coded aperture computed tomography (CT) is proposed in this paper to alleviate the tradeoff between the non-continuous sparse projections and the ill-posedness iterative reconstruction problem. A non-continuous sparse projection model is presented based on generative adversarial U-net and the corresponding joint penalty function is formulated. Simulations using real datasets show that CT images with 256×256 pixels can be reconstructed with peak signal-to-noise ration more than 30 dB at only 5% transmittance. Furthermore, the computational time in the reconstructions is reduced by two orders of magnitude when compared with the state-of-the-art iterative algorithms in coded aperture computed tomography.
2022, 31(3): 329-336.
doi: 10.15953/j.ctta.2021.005
Abstract:
Objective: To explore the application of deep learning image reconstruction (DLIR) algorithm in upper abdominal CT imaging by analyzing the image quality of adaptive statistical iterative reconstruction (ASIR) algorithm and DLIR. Methods: Retrospectively included 75 patients’ upper abdominal CT plain scan images, using adaptive statistical iterative reconstruction algorithm ASIR (30%, 50%, 70%, 90%) and deep learning reconstruction algorithm (DL-L, DL-M, DL-H) to reconstruct images, a total of 7 groups. Measured the CT and SD values of the liver, pancreas, and erector spinae , and calculated the signal to noise ratio (SNR) and contrast to noise ratio (CNR). Objective indicators were evaluated by one-way ANOVA. Two radiologists scored the image quality and noise, and compared them with Friedman M test. Results: (1) The SD value, SNR, and liver CNR of the seven reconstructed images had statistically significant differences. (2) The difference in CT value, SD value, SNR value and CNR value at each ROI between DL-L and ASIR 50%, DL-M and ASIR 70%, DL-H and ASIR 90% was small. (3) The SNR value of the three DLIR algorithms increased as the level increased, and the difference was statistically significant; and the SNR value of the DL-H algorithm was higher than ASIR 30% and ASIR 50%, and the SD value was lower than the other five reconstruction algorithms except for the ASIR 90%. (4) The difference in the subjective scores of the seven groups of images was statistically significant. The algorithm DL-H had the best image quality and the lowest noise, DL-M, ASIR 90%, DL-L, ASIR 70%, ASIR 50%, ASIR 30% image noise in sequence increased. Conclusion: The DLIR algorithm can reduce the image noise of the upper abdomen and improve the image quality. As the level increased, the image noise decreased, the quality improved, and the signal-to-noise ratio increased.
Objective: To explore the application of deep learning image reconstruction (DLIR) algorithm in upper abdominal CT imaging by analyzing the image quality of adaptive statistical iterative reconstruction (ASIR) algorithm and DLIR. Methods: Retrospectively included 75 patients’ upper abdominal CT plain scan images, using adaptive statistical iterative reconstruction algorithm ASIR (30%, 50%, 70%, 90%) and deep learning reconstruction algorithm (DL-L, DL-M, DL-H) to reconstruct images, a total of 7 groups. Measured the CT and SD values of the liver, pancreas, and erector spinae , and calculated the signal to noise ratio (SNR) and contrast to noise ratio (CNR). Objective indicators were evaluated by one-way ANOVA. Two radiologists scored the image quality and noise, and compared them with Friedman M test. Results: (1) The SD value, SNR, and liver CNR of the seven reconstructed images had statistically significant differences. (2) The difference in CT value, SD value, SNR value and CNR value at each ROI between DL-L and ASIR 50%, DL-M and ASIR 70%, DL-H and ASIR 90% was small. (3) The SNR value of the three DLIR algorithms increased as the level increased, and the difference was statistically significant; and the SNR value of the DL-H algorithm was higher than ASIR 30% and ASIR 50%, and the SD value was lower than the other five reconstruction algorithms except for the ASIR 90%. (4) The difference in the subjective scores of the seven groups of images was statistically significant. The algorithm DL-H had the best image quality and the lowest noise, DL-M, ASIR 90%, DL-L, ASIR 70%, ASIR 50%, ASIR 30% image noise in sequence increased. Conclusion: The DLIR algorithm can reduce the image noise of the upper abdomen and improve the image quality. As the level increased, the image noise decreased, the quality improved, and the signal-to-noise ratio increased.
