Jing Wang
Jing Wang has a total of 24 co-authors. Here are the top co-authors by number of shared publications:
- Chuxin Wang
School of Information Science and Technology, University of Science and Technology of China (USTC), Hefei, China
1 shared publication - Dongsheng Zhang
School of Mechanical Engineering, Inner Mongolia University of Technology, Hohhot, China
1 shared publication - Fei Cao
Key Laboratory of RF Circuits and Systems, Ministry of Education, Hangzhou Dianzi University, Hangzhou, China
1 shared publication - Fengming Jin
Dagang Oilfield Company, China National Petroleum Corporation, Tianjin, China
1 shared publication - Haishan Luo
TOTAL
1 shared publication - Hongling Zhang
Henan Key Laboratory of Nanomedicine for Targeting Diagnosis and Treatment, School of Pharmaceutical Sciences, Zhengzhou University, Zhengzhou, 450001, China
1 shared publication
Latest Publications
- Open Access
- 21 November 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 21 (12) , 1-23
ER-Depth: Enhancing the Robustness of Self-Supervised Monocular Depth Estimation in Challenging Scenes
Self-supervised monocular depth estimation holds significant importance in the fields of autonomous driving and robotics. However, existing methods are typically trained and evaluated on clear, sunny datasets, overlooking the impact of various adverse conditions commonly encountered in real-world applications, such as rainy weather, low visibility, and motion blur. As a result, they often struggle in challenging scenarios and produce artifacts. To address this issue, we propose ER-Depth, a novel two-stage self-supervised framework designed for robust depth estimation. In the first stage, we propose perturbation-invariant depth consistency regularization to propagate reliable supervision from standard to challenging scenes. In the second stage, we adopt the Mean Teacher paradigm for self-distillation and present a novel consistency-based pseudo-label filtering strategy to improve the quality of pseudo-labels. Extensive experiments demonstrate that our method exhibits exceptional robustness in challenging scenarios while maintaining high performance in standard scenes, significantly outperforming existing state-of-the-art methods on challenging KITTI-C, DrivingStereo, and NuScenes-Night benchmarks. Project page: https://ruijiezhu94.github.io/ERDepth_page .
- 21 November 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 21 (12) , 1-23
Open AccessER-Depth: Enhancing the Robustness of Self-Supervised Monocular Depth Estimation in Challenging Scenes
Self-supervised monocular depth estimation holds significant importance in the fields of autonomous driving and robotics. However, existing methods are typically trained and evaluated on clear, sunny datasets, overlooking the impact of various adverse conditions commonly encountered in real-world applications, such as rainy weather, low visibility, and motion blur. As a result, they often struggle in challenging scenarios and produce artifacts. To address this issue, we propose ER-Depth, a novel two-stage self-supervised framework designed for robust depth estimation. In the first stage, we propose perturbation-invariant depth consistency regularization to propagate reliable supervision from standard to challenging scenes. In the second stage, we adopt the Mean Teacher paradigm for self-distillation and present a novel consistency-based pseudo-label filtering strategy to improve the quality of pseudo-labels. Extensive experiments demonstrate that our method exhibits exceptional robustness in challenging scenarios while maintaining high performance in standard scenes, significantly outperforming existing state-of-the-art methods on challenging KITTI-C, DrivingStereo, and NuScenes-Night benchmarks. Project page: https://ruijiezhu94.github.io/ERDepth_page .
- Open Access
- 20 June 2020
- Published by Wiley in Shock and Vibration
- Vol. 2020, 1-15
Fault Diagnosis of Gearbox in Multiple Conditions Based on Fine-Grained Classification CNN Algorithm
The use of the convolutional neural network for fault diagnosis has been a common method of research in recent years. Since this method can automatically extract fault features, it has played a good role in some research studies. However, this method has a clear drawback that the signals will be significantly affected by working conditions and sample size, and it is difficult to improve diagnostic accuracy by directly learning faults, regardless of working conditions. It is therefore a research orientation worthy of a diagnosis of high precision defect in various working conditions. In this article, using a fine-grained classification algorithm, the operating conditions of the object system are considered an approximate classification. A specific failure in different working conditions is considered a beautiful classification. Samples of different faults in different working conditions are learned uniformly and the common characteristics are extracted from the convolutional network so that different faults of different working conditions can simultaneously be identified on the basis of the entire sample. Experimental results show that the method effectively uses the set of samples of the working conditions of the variables to obtain the dual recognition of defects and specific working conditions and the accuracy of the recognition is significantly higher than the method of learning regardless of working conditions.
- 20 June 2020
- Published by Wiley in Shock and Vibration
- Vol. 2020, 1-15
Open AccessFault Diagnosis of Gearbox in Multiple Conditions Based on Fine-Grained Classification CNN Algorithm
The use of the convolutional neural network for fault diagnosis has been a common method of research in recent years. Since this method can automatically extract fault features, it has played a good role in some research studies. However, this method has a clear drawback that the signals will be significantly affected by working conditions and sample size, and it is difficult to improve diagnostic accuracy by directly learning faults, regardless of working conditions. It is therefore a research orientation worthy of a diagnosis of high precision defect in various working conditions. In this article, using a fine-grained classification algorithm, the operating conditions of the object system are considered an approximate classification. A specific failure in different working conditions is considered a beautiful classification. Samples of different faults in different working conditions are learned uniformly and the common characteristics are extracted from the convolutional network so that different faults of different working conditions can simultaneously be identified on the basis of the entire sample. Experimental results show that the method effectively uses the set of samples of the working conditions of the variables to obtain the dual recognition of defects and specific working conditions and the accuracy of the recognition is significantly higher than the method of learning regardless of working conditions.
- 26 September 2016
- Published by Society of Petroleum Engineers (SPE)
How to Correct the Petro-Physical Properties for Simulating Shale Gas Production Using Current Commercial Simulators?
Numerical simulation is important to understand and predict the development of oil and gas reservoirs. Existing commercial simulators, such as CMG, ECLIPSE and VIP, have been widely used in the past several decades for their robust performance in computing and scaling. In these reservoir simulators, the fluid flow models are based on Darcy's law or its extended form; the volume of the adsorbed phase or component is overlooked as well. While this is acceptable for conventional oil/gas reservoirs or chemical flooding reservoirs, the gas flow regimes such as slippage flow, transition flow, and molecular-free flow significantly is deviated from Darcy flow for shale gas reservoirs. Besides, a large portion of the gas is stored in the pore in the form of adsorbed gas. If the volume of the adsorbed gas is still overlooked, the volume of free gas and original gas in place (OGIP) will be seriously overestimated. For the above reasons, it is commonly thought that existing commercial simulators could not be ideally used to simulate the development of shale gas reservoirs. Hence, it is desirable to attain a feasible approach to correct the petro-physical properties of shale gas effectively within the commercial simulators, in order that one can use them to accurately simulate the development of shale gas reservoirs. In this paper, we first derived the correction formulas for the bulk porosity, free gas saturation, and connate water saturation used for correcting the disregarded volume of adsorbed gas in commercial simulators. Then, we derived the models of permeability and porosity multipliers in matrix considering gas adsorption/desorption, geomechanics, non-Darcy flow regimes, and diffusion of adsorbed layer. Finally, the above models were used to attain the corrected petro-physical properties for simulating gas production based on the practical properties of shale gas reservoirs using commercial simulators. The validation was performed by comparing the simulation results of commercial simulator with the published mechanism simulator using gas field data. The results show that the simulation results using commercial simulator achieve good agreement with the published mechanism simulator with the corrected petro-physical properties. The corrections of bulk porosity, connate water saturation, and free gas saturation are very essential. The correction formulas for these properties can largely decreases the error of OGIP and the calculated gas production. Both permeability and porosity multipliers are the functions of gas pressure, but they are not a monotone decreasing/increasing function. The gas production may be significantly overestimated or underestimated without consideration of these characteristics of shale gas in different fields. The contributions of different mechanisms are also demonstrated using the commercial simulator. This work can evently solve the issue that existing commercial simulators cannot accurately simulate shale gas production. The researchers can easily use these commercial simulators with these corrected formulas, which is a great progress for modeling the development of shale gas reservoir.
- 26 September 2016
- Published by Society of Petroleum Engineers (SPE)
How to Correct the Petro-Physical Properties for Simulating Shale Gas Production Using Current Commercial Simulators?
Numerical simulation is important to understand and predict the development of oil and gas reservoirs. Existing commercial simulators, such as CMG, ECLIPSE and VIP, have been widely used in the past several decades for their robust performance in computing and scaling. In these reservoir simulators, the fluid flow models are based on Darcy's law or its extended form; the volume of the adsorbed phase or component is overlooked as well. While this is acceptable for conventional oil/gas reservoirs or chemical flooding reservoirs, the gas flow regimes such as slippage flow, transition flow, and molecular-free flow significantly is deviated from Darcy flow for shale gas reservoirs. Besides, a large portion of the gas is stored in the pore in the form of adsorbed gas. If the volume of the adsorbed gas is still overlooked, the volume of free gas and original gas in place (OGIP) will be seriously overestimated. For the above reasons, it is commonly thought that existing commercial simulators could not be ideally used to simulate the development of shale gas reservoirs. Hence, it is desirable to attain a feasible approach to correct the petro-physical properties of shale gas effectively within the commercial simulators, in order that one can use them to accurately simulate the development of shale gas reservoirs. In this paper, we first derived the correction formulas for the bulk porosity, free gas saturation, and connate water saturation used for correcting the disregarded volume of adsorbed gas in commercial simulators. Then, we derived the models of permeability and porosity multipliers in matrix considering gas adsorption/desorption, geomechanics, non-Darcy flow regimes, and diffusion of adsorbed layer. Finally, the above models were used to attain the corrected petro-physical properties for simulating gas production based on the practical properties of shale gas reservoirs using commercial simulators. The validation was performed by comparing the simulation results of commercial simulator with the published mechanism simulator using gas field data. The results show that the simulation results using commercial simulator achieve good agreement with the published mechanism simulator with the corrected petro-physical properties. The corrections of bulk porosity, connate water saturation, and free gas saturation are very essential. The correction formulas for these properties can largely decreases the error of OGIP and the calculated gas production. Both permeability and porosity multipliers are the functions of gas pressure, but they are not a monotone decreasing/increasing function. The gas production may be significantly overestimated or underestimated without consideration of these characteristics of shale gas in different fields. The contributions of different mechanisms are also demonstrated using the commercial simulator. This work can evently solve the issue that existing commercial simulators cannot accurately simulate shale gas production. The researchers can easily use these commercial simulators with these corrected formulas, which is a great progress for modeling the development of shale gas reservoir.
- 25 September 2014
- Published by Wiley in Journal of Petroleum Geology
- Vol. 37 (4) , 391-404
GENERATION AND ACCUMULATION OF HYDROCARBONS IN A DEEP “BURIED HILL” STRUCTURE IN THE BAXIAN DEPRESSION, BOHAI BAY BASIN, EASTERN CHINA
Significant volumes of hydrocarbons have been produced from karstified Infracambrian dolomites in a “buried hill” structure at depths of 5860m to 6027m and reservoir temperatures of 190–201°C in well Niudong‐1 in the Baxian depression, Bohai Bay Basin. This is the deepest oil and gas discovery made in eastern China so far and structures at similar depths are targets for exploration elsewhere in the Bohai Bay Basin. However the origin and accumulation of the hydrocarbons at Niudong‐1 is not clear: they may have been generated from highly mature lacustrine source rocks in the Sha‐4 Member of the Eocene Shahejie Formation; or they may have been derived from thermal cracking of previously‐accumulated oil.This paper investigates the organic geochemistry of the Sha‐4 Member source rocks and the crude oils produced from well Niudong‐1. Analyses of molecular parameters show that the hydrocarbons originated from the pyrolysis of organic matter in Sha‐4 Member source rocks, rather than from cracking of previously accumulated oil. Infracambrian dolomites at the Niudong‐1 location may have been charged with low‐maturity oil around 34 Ma ago, when the Sha‐4 Member source rocks were buried to depths of about 3500m and first entered the oil window. During further rapid burial to more than 5500m starting at about 15Ma, these source rocks became highly mature and generated significant volumes of light oil and gas. Overpressures in the source rock interval forced these hydrocarbons to migrate into unconformably‐underlying Infracambrian dolomite reservoir rocks at the Niudong‐1 structure.Significant risks are associated with future exploration of deep “buried hill” structures in the Bohai Bay Basin. Not all the structures were charged with oil, and accumulations were not necessarily preserved during Neogene burial as the reservoirs may have been breached by faulting.
- 25 September 2014
- Published by Wiley in Journal of Petroleum Geology
- Vol. 37 (4) , 391-404
GENERATION AND ACCUMULATION OF HYDROCARBONS IN A DEEP “BURIED HILL” STRUCTURE IN THE BAXIAN DEPRESSION, BOHAI BAY BASIN, EASTERN CHINA
Significant volumes of hydrocarbons have been produced from karstified Infracambrian dolomites in a “buried hill” structure at depths of 5860m to 6027m and reservoir temperatures of 190–201°C in well Niudong‐1 in the Baxian depression, Bohai Bay Basin. This is the deepest oil and gas discovery made in eastern China so far and structures at similar depths are targets for exploration elsewhere in the Bohai Bay Basin. However the origin and accumulation of the hydrocarbons at Niudong‐1 is not clear: they may have been generated from highly mature lacustrine source rocks in the Sha‐4 Member of the Eocene Shahejie Formation; or they may have been derived from thermal cracking of previously‐accumulated oil.This paper investigates the organic geochemistry of the Sha‐4 Member source rocks and the crude oils produced from well Niudong‐1. Analyses of molecular parameters show that the hydrocarbons originated from the pyrolysis of organic matter in Sha‐4 Member source rocks, rather than from cracking of previously accumulated oil. Infracambrian dolomites at the Niudong‐1 location may have been charged with low‐maturity oil around 34 Ma ago, when the Sha‐4 Member source rocks were buried to depths of about 3500m and first entered the oil window. During further rapid burial to more than 5500m starting at about 15Ma, these source rocks became highly mature and generated significant volumes of light oil and gas. Overpressures in the source rock interval forced these hydrocarbons to migrate into unconformably‐underlying Infracambrian dolomite reservoir rocks at the Niudong‐1 structure.Significant risks are associated with future exploration of deep “buried hill” structures in the Bohai Bay Basin. Not all the structures were charged with oil, and accumulations were not necessarily preserved during Neogene burial as the reservoirs may have been breached by faulting.
- 3 April 2008
- Published by SPIE-Intl Soc Optical Eng
- Vol. 6969, 696905-696905-8
Detection of small objects in multi-layered infrared images
This paper uses super-resolution methods to detect small objects in infrared image sequences from a simulated airborne platform, using image registration techniques for automatic sightline stabilisation. The scene consists of multiple layers, corresponding to a static background scene and layers of cloud cover at varying heights. The motivation is to evaluate the performance of super-resolution methods in the presence of three-dimensional structured infrared clutter.
- 3 April 2008
- Published by SPIE-Intl Soc Optical Eng
- Vol. 6969, 696905-696905-8
Detection of small objects in multi-layered infrared images
This paper uses super-resolution methods to detect small objects in infrared image sequences from a simulated airborne platform, using image registration techniques for automatic sightline stabilisation. The scene consists of multiple layers, corresponding to a static background scene and layers of cloud cover at varying heights. The motivation is to evaluate the performance of super-resolution methods in the presence of three-dimensional structured infrared clutter.