Yingtian Hu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Yingtian Hu has a total of 48 co-authors. Here are the top co-authors by number of shared publications:
- Dongdong Zhao
College of Computer Science and Technology, Zhejiang University of Technology, Hangzhou 310023, China
11 shared publications - Peng Chen
College of Computer Science and Technology, Zhejiang University of Technology, Hangzhou, China
6 shared publications - Ronghua Liang
College of Computer Science and Technology, Zhejiang University of Technology, Hangzhou, China
6 shared publications - Yuanjie Dang
Zhejiang University of Technology
5 shared publications - Huan Zheng
Institute of Fiber-optic Communication and Information Engineering, College of Information Engineering, Zhejiang University of Technology, Hangzhou 310023, China
4 shared publications - Xinxin Guo
Institute of Deep-Sea Science and Engineering, Chinese Academy of Sciences, Sanya, Hainan, China
4 shared publications
Latest Publications
- 16 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13  (7) , 13095-13109
Design of a Terminal Sensor With Super-Resolution Reconstruction Algorithm Used in IoT-Based Water Quality Monitoring
In internet of things (IoT)-based water quality monitoring, traditional 3-D fluorescence spectroscopy is difficult to deploy directly on terminal sensors due to its large size, high cost, and complex structure, despite its rich information being highly advantageous for water pollution detection. In this work, a novel approach is proposed through designing a compact sensor integrated with a super-resolution reconstruction algorithm to meet the requirement of entire 3-D fluorescence spectra in cloud deployment of monitoring network at minimal cost. In the design of the terminal sensor, multiple light-emitting diodes (LEDs) with different wavelengths serve as the excitation light source and multiple filters are used to switch emission wavelengths, enabling the acquisition of a sparse 3-D fluorescence spectrum. In the reconstruction algorithm, a point spread function (PSF) is employed as a constraint module to simulate the degradation process of spectra from high to low resolution by modeling the spectral characteristics of the optical components responsible for degradation, including LEDs and optical filters, which can enhance the authenticity of the reconstructed spectra. Furthermore, a joint training strategy is introduced for better robustness by jointly optimizing the parameters of PSF and deep learning network. Experimental results demonstrate that the proposed method achieves the highest reconstruction quality peak signal-to-noise ratio (PSNR): 47.57 dB, structural similarity (SSIM): 0.9903, and pollutant identification accuracy (98.1%) compared with seven other reconstruction methods. The proposed approach can obtain high-resolution (HR) and entire 3-D fluorescence spectra for water pollution detection with low manufacturing cost and minimal data, thereby meeting the practical application requirements in the field of IoT-based water quality monitoring.
- 16 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13  (7) , 13095-13109
Design of a Terminal Sensor With Super-Resolution Reconstruction Algorithm Used in IoT-Based Water Quality Monitoring
In internet of things (IoT)-based water quality monitoring, traditional 3-D fluorescence spectroscopy is difficult to deploy directly on terminal sensors due to its large size, high cost, and complex structure, despite its rich information being highly advantageous for water pollution detection. In this work, a novel approach is proposed through designing a compact sensor integrated with a super-resolution reconstruction algorithm to meet the requirement of entire 3-D fluorescence spectra in cloud deployment of monitoring network at minimal cost. In the design of the terminal sensor, multiple light-emitting diodes (LEDs) with different wavelengths serve as the excitation light source and multiple filters are used to switch emission wavelengths, enabling the acquisition of a sparse 3-D fluorescence spectrum. In the reconstruction algorithm, a point spread function (PSF) is employed as a constraint module to simulate the degradation process of spectra from high to low resolution by modeling the spectral characteristics of the optical components responsible for degradation, including LEDs and optical filters, which can enhance the authenticity of the reconstructed spectra. Furthermore, a joint training strategy is introduced for better robustness by jointly optimizing the parameters of PSF and deep learning network. Experimental results demonstrate that the proposed method achieves the highest reconstruction quality peak signal-to-noise ratio (PSNR): 47.57 dB, structural similarity (SSIM): 0.9903, and pollutant identification accuracy (98.1%) compared with seven other reconstruction methods. The proposed approach can obtain high-resolution (HR) and entire 3-D fluorescence spectra for water pollution detection with low manufacturing cost and minimal data, thereby meeting the practical application requirements in the field of IoT-based water quality monitoring.
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Photonics Technology Letters
- Vol. PP  (10411135) , 1
Mixed Contaminants Identification Based on 3D Fluorescent Spectroscopy with TTE Model
In this letter, a method of mixed contaminants identification based on 3D fluorescent spectroscopy with a novel generative model is proposed. To comprehensively characterize the 3D fluorescent spectra of contaminants, local and global features can be respectively extracted to represent the fluorescent peaks of substances and the overall trend of 3D fluorescent spectroscopy. Local features involving substance properties can be beneficial for improving authenticity, whereas global features can increase diversity as auxiliary information. For that, a Text-to-EEM (TTE) generative model is presented with convolution and retentive self-attention mechanisms to capture local and global features, respectively. In addition, spectral data can be generated from textual descriptions using text generation modules. Experiments using CNN identification network show that the maximum accuracy of samples generated by the TTE model was 67% for all correct and 20% for partially correct, better than GAN, WGAN, and WAE.
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Photonics Technology Letters
- Vol. PP  (10411135) , 1
Mixed Contaminants Identification Based on 3D Fluorescent Spectroscopy with TTE Model
In this letter, a method of mixed contaminants identification based on 3D fluorescent spectroscopy with a novel generative model is proposed. To comprehensively characterize the 3D fluorescent spectra of contaminants, local and global features can be respectively extracted to represent the fluorescent peaks of substances and the overall trend of 3D fluorescent spectroscopy. Local features involving substance properties can be beneficial for improving authenticity, whereas global features can increase diversity as auxiliary information. For that, a Text-to-EEM (TTE) generative model is presented with convolution and retentive self-attention mechanisms to capture local and global features, respectively. In addition, spectral data can be generated from textual descriptions using text generation modules. Experiments using CNN identification network show that the maximum accuracy of samples generated by the TTE model was 67% for all correct and 20% for partially correct, better than GAN, WGAN, and WAE.
- 6 November 2025
- Published by Springer Nature in Journal of Applied Spectroscopy
- Vol. 92  (5) , 1162-1171
Spectral Decomposition Method for Water Chemical Oxygen Demand Measurement Based on UV-Vis Absorption Spectroscopy
- 6 November 2025
- Published by Springer Nature in Journal of Applied Spectroscopy
- Vol. 92  (5) , 1162-1171
Spectral Decomposition Method for Water Chemical Oxygen Demand Measurement Based on UV-Vis Absorption Spectroscopy
- 12 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 35  (10) , 9695-9707
SonarPoint: Weak-Heterogeneity Awareness Object Detection Network for 3D Sonar Point Cloud
Underwater target detection is primarily achieved through two methods: optical imaging and underwater sonar. 3D sonar, as the most advanced underwater detection technology, is characterized by strong penetration and long scanning distance, making it more suitable for tasks such as deep-sea exploration, murky water detection, and long-distance target identification. However, acquiring underwater sonar images is challenging, and there is no open-source 3D sonar dataset. Traditional three-dimensional target detection methods typically require highquality data and face significant challenges when dealing with weak heterogeneous sonar point clouds caused by high noise, low resolution, and occlusions. To address the aforementioned issues, we first propose a novel fuzzy decoupling module that differs from traditional foreground-background segmentation. This module simultaneously extracts valuable information about the target and its surrounding environment, mitigating the reduction in heterogeneity caused by noise and sonar side lobes. To achieve efficient fusion and capture global information after fuzzy decoupling, a multi-hop Mamba seamless adaptive decoupling point is introduced. It effectively enhances the connection between the two decoupled parts. To address missing and occlusion problems, a second-stage refinement based on Markov prediction is proposed. This low-cost approach, in contrast to using the original point cloud for contour and detail completion, enriches target boundary information. To validate our method, we have designed a practical 3D sonar imaging system and tested it through lake-based experiments. We have collected extensive raw data from Qiandao Lake and conducted annotation work. Through qualitative and quantitative experiments, our method outperforms the most advanced methods by 11.4%.
- 12 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 35  (10) , 9695-9707
SonarPoint: Weak-Heterogeneity Awareness Object Detection Network for 3D Sonar Point Cloud
Underwater target detection is primarily achieved through two methods: optical imaging and underwater sonar. 3D sonar, as the most advanced underwater detection technology, is characterized by strong penetration and long scanning distance, making it more suitable for tasks such as deep-sea exploration, murky water detection, and long-distance target identification. However, acquiring underwater sonar images is challenging, and there is no open-source 3D sonar dataset. Traditional three-dimensional target detection methods typically require highquality data and face significant challenges when dealing with weak heterogeneous sonar point clouds caused by high noise, low resolution, and occlusions. To address the aforementioned issues, we first propose a novel fuzzy decoupling module that differs from traditional foreground-background segmentation. This module simultaneously extracts valuable information about the target and its surrounding environment, mitigating the reduction in heterogeneity caused by noise and sonar side lobes. To achieve efficient fusion and capture global information after fuzzy decoupling, a multi-hop Mamba seamless adaptive decoupling point is introduced. It effectively enhances the connection between the two decoupled parts. To address missing and occlusion problems, a second-stage refinement based on Markov prediction is proposed. This low-cost approach, in contrast to using the original point cloud for contour and detail completion, enriches target boundary information. To validate our method, we have designed a practical 3D sonar imaging system and tested it through lake-based experiments. We have collected extensive raw data from Qiandao Lake and conducted annotation work. Through qualitative and quantitative experiments, our method outperforms the most advanced methods by 11.4%.
- 21 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74  (00189456) , 1-15
Design of Deep-Sea Imaging Sonar System With Universal-Scene Sparse Array
Imaging sonar plays a critical role in deep-sea exploration and surveying tasks, with miniaturization and high precision being the developmental trends. Traditional designs often focus on adopting sparse arrays to reduce system complexity, but fail to consider the challenges posed by non-ideal underwater environments, especially the degradation of sonar images caused by varying sound velocities and defocusing, along with the harsh requirements for hardware design imposed by extremely high water pressures. To address this issue, we propose a novel design of deep-sea imaging sonar system with universal-scene sparse array (USSA). At the algorithmic level, we propose the USSA and introduce a novel simulated annealing energy function that enhances beamforming performance in both far-field and near-field conditions across a wide range of sound velocities, achieving at least a 3.21 dB reduction in the highest sidelobe in the near field compared to traditional methods. Simultaneously, we propose a velocity-based direct image correction method, which can directly rectify sonar image distortions caused by mismatches between the actual and expected sound velocities, and the structural similarity index measure (SSIM) is improved by at least 5.1%. In terms of hardware, to accommodate the algorithm’s pursuit of universal-scene applicability, we propose a field-programmable gate array (FPGA)-based hybrid system architecture and a pressuretolerant design. This hardware system features reconfigurability and enhanced universal-scene adaptability, providing a reliable platform for algorithm implementation. Based on these concepts, we fabricated a prototype. During a sea trial in the South China Sea, the prototype mounted on a deep-sea crawler and deployed at a depth of 4098 meters under the sea, produced satisfactory images.
- 21 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74  (00189456) , 1-15
Design of Deep-Sea Imaging Sonar System With Universal-Scene Sparse Array
Imaging sonar plays a critical role in deep-sea exploration and surveying tasks, with miniaturization and high precision being the developmental trends. Traditional designs often focus on adopting sparse arrays to reduce system complexity, but fail to consider the challenges posed by non-ideal underwater environments, especially the degradation of sonar images caused by varying sound velocities and defocusing, along with the harsh requirements for hardware design imposed by extremely high water pressures. To address this issue, we propose a novel design of deep-sea imaging sonar system with universal-scene sparse array (USSA). At the algorithmic level, we propose the USSA and introduce a novel simulated annealing energy function that enhances beamforming performance in both far-field and near-field conditions across a wide range of sound velocities, achieving at least a 3.21 dB reduction in the highest sidelobe in the near field compared to traditional methods. Simultaneously, we propose a velocity-based direct image correction method, which can directly rectify sonar image distortions caused by mismatches between the actual and expected sound velocities, and the structural similarity index measure (SSIM) is improved by at least 5.1%. In terms of hardware, to accommodate the algorithm’s pursuit of universal-scene applicability, we propose a field-programmable gate array (FPGA)-based hybrid system architecture and a pressuretolerant design. This hardware system features reconfigurability and enhanced universal-scene adaptability, providing a reliable platform for algorithm implementation. Based on these concepts, we fabricated a prototype. During a sea trial in the South China Sea, the prototype mounted on a deep-sea crawler and deployed at a depth of 4098 meters under the sea, produced satisfactory images.
- 5 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73  (00189456) , 1-9
Design of Rapid Water Contaminant Identification System Based on 3-D Topological Structure of Fluorescent Excitation–Emission Matrix
Existing water contaminant identification systems based on emission–emission matrix (EEM) mostly focus on the data level using a 2-D matrix, which makes it difficult to distinguish the EEM with similar location and shape owing to the insensitivity of a matrix in contrast to the 3-D morphology. In this study, a system of water contaminant identification was presented with an innovative approach based on the theory of 3-D topology and the EEM-oriented Reeb graph was proposed as a new spectral feature. In this system, an automatic scanning system for EEM was designed and constructed, and relevant procedures of analysis algorithms were integrated in it. The EEM-oriented Reeb graph describes the key 3-D morphological characteristics of fluorescence peaks without redundant information so that they are sensitive to the change of spatial structure and benefit for distinguishment. In addition, owing to topological invariance, the EEM-oriented Reeb graph is insensitive to factors, such as concentration and peak shift, and is therefore more robust for substance identification. The new method for water contaminant identification was assessed through controlled laboratory experiments for simulating the presence of six typical contaminants commonly found in the river. The samples containing multiple pollutants were identified, and the accuracy was 75% all correct and 16.7% partial correct better than principal component analysis (PCA), wavelet decomposition, and conventional peak picking. In addition, the time consumed in constructing models was only around 23 ms, shorter than PCA and wavelet decomposition.
- 5 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73  (00189456) , 1-9
Design of Rapid Water Contaminant Identification System Based on 3-D Topological Structure of Fluorescent Excitation–Emission Matrix
Existing water contaminant identification systems based on emission–emission matrix (EEM) mostly focus on the data level using a 2-D matrix, which makes it difficult to distinguish the EEM with similar location and shape owing to the insensitivity of a matrix in contrast to the 3-D morphology. In this study, a system of water contaminant identification was presented with an innovative approach based on the theory of 3-D topology and the EEM-oriented Reeb graph was proposed as a new spectral feature. In this system, an automatic scanning system for EEM was designed and constructed, and relevant procedures of analysis algorithms were integrated in it. The EEM-oriented Reeb graph describes the key 3-D morphological characteristics of fluorescence peaks without redundant information so that they are sensitive to the change of spatial structure and benefit for distinguishment. In addition, owing to topological invariance, the EEM-oriented Reeb graph is insensitive to factors, such as concentration and peak shift, and is therefore more robust for substance identification. The new method for water contaminant identification was assessed through controlled laboratory experiments for simulating the presence of six typical contaminants commonly found in the river. The samples containing multiple pollutants were identified, and the accuracy was 75% all correct and 16.7% partial correct better than principal component analysis (PCA), wavelet decomposition, and conventional peak picking. In addition, the time consumed in constructing models was only around 23 ms, shorter than PCA and wavelet decomposition.
- 14 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73  (00189456) , 1-17
Design of Forward-Looking Sonar System for Real-Time Image Segmentation With Light Multiscale Attention Net
Forward-looking sonar is a commonly used underwater detection device. However, due to the complex underwater environment, small target areas, and blurred features, the detection accuracy is poor and not conducive to quickly finding the desired targets. In recent years, deep learning has developed rapidly, and semantic segmentation network models have shown strong segmentation performance, with enormous application potential in forward-looking sonar image segmentation. A lightweight multiscale attention network is proposed in this article for real-time semantic segmentation of forward-looking sonar system. The proposed model ensures the training accuracy and inference speed of the semantic segmentation network through the structural reparameterization module and improves the feature extraction ability of the network by fusing multiscale information through the multiscale feature attention gate. Compared with other models, the proposed model achieved good results on semantic segmentation datasets and data collected from self-developed devices. The mean intersection over union (mIoU) reached 0.734 on a public dataset, outperforming existing mainstream models. Moreover, the inference speed can reach 1250 frame rate per second (FPS) on 2080Ti. An embedded forward-looking sonar host computer is designed based on Atlas200, and the proposed network can run up to 87 FPS on the embedded host computer, with a frame rate of 9.5 FPS per unit of power consumption, which is better than that of mobile phones and X86 computers. The forward-looking sonar system constructed based on this achieved satisfactory results in engineering practice.
- 14 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73  (00189456) , 1-17
Design of Forward-Looking Sonar System for Real-Time Image Segmentation With Light Multiscale Attention Net
Forward-looking sonar is a commonly used underwater detection device. However, due to the complex underwater environment, small target areas, and blurred features, the detection accuracy is poor and not conducive to quickly finding the desired targets. In recent years, deep learning has developed rapidly, and semantic segmentation network models have shown strong segmentation performance, with enormous application potential in forward-looking sonar image segmentation. A lightweight multiscale attention network is proposed in this article for real-time semantic segmentation of forward-looking sonar system. The proposed model ensures the training accuracy and inference speed of the semantic segmentation network through the structural reparameterization module and improves the feature extraction ability of the network by fusing multiscale information through the multiscale feature attention gate. Compared with other models, the proposed model achieved good results on semantic segmentation datasets and data collected from self-developed devices. The mean intersection over union (mIoU) reached 0.734 on a public dataset, outperforming existing mainstream models. Moreover, the inference speed can reach 1250 frame rate per second (FPS) on 2080Ti. An embedded forward-looking sonar host computer is designed based on Atlas200, and the proposed network can run up to 87 FPS on the embedded host computer, with a frame rate of 9.5 FPS per unit of power consumption, which is better than that of mobile phones and X86 computers. The forward-looking sonar system constructed based on this achieved satisfactory results in engineering practice.
- Top Cited Papers
- 7 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 43  (5) , 1401-1414
A Distributed and Parallel Accelerator Design for 3-D Acoustic Imaging on FPGA-Based Systems
3-D imaging sonar is crucial in the exploration of marine resources, and the development of portable device with high-imaging quality and high-real-time performance is the general trend. However, traditional framework methods are limited by the huge amount of computation brought by high-quality imaging, making it difficult to implement in engineering. To address this issue, we develop 3-D real-time sonar system in an algorithm-hardware co-designed way. An ultrawideband distributed and parallel subarray beamforming algorithm (UWB-DPS) is proposed for 3-D acoustic imaging. This is a multistage array time-frequency beamforming method under a distributed parallel computing architecture. Based on this, we propose field-programmable gate array (FPGA)-based accelerator. It divides a large sonar receiving planar array into multiple parallel subarrays, and complete the beamforming in two stages, which can reduce the calculation load and speeds up 3-D imaging. For engineering implementation, we optimized the sparseness of the planar transducer array, with a sparse rate as high as 97.7%. The experimental results show that the calculation amount of the proposed UWB-DPS algorithm is reduced to 1/5.7 of the traditional framework algorithm, the imaging performance is effectively improved, and the FPGA-based accelerator outperforms the CPU software implementation by 935 $\times $ .
- 7 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 43  (5) , 1401-1414
Top Cited PapersA Distributed and Parallel Accelerator Design for 3-D Acoustic Imaging on FPGA-Based Systems
3-D imaging sonar is crucial in the exploration of marine resources, and the development of portable device with high-imaging quality and high-real-time performance is the general trend. However, traditional framework methods are limited by the huge amount of computation brought by high-quality imaging, making it difficult to implement in engineering. To address this issue, we develop 3-D real-time sonar system in an algorithm-hardware co-designed way. An ultrawideband distributed and parallel subarray beamforming algorithm (UWB-DPS) is proposed for 3-D acoustic imaging. This is a multistage array time-frequency beamforming method under a distributed parallel computing architecture. Based on this, we propose field-programmable gate array (FPGA)-based accelerator. It divides a large sonar receiving planar array into multiple parallel subarrays, and complete the beamforming in two stages, which can reduce the calculation load and speeds up 3-D imaging. For engineering implementation, we optimized the sparseness of the planar transducer array, with a sparse rate as high as 97.7%. The experimental results show that the calculation amount of the proposed UWB-DPS algorithm is reduced to 1/5.7 of the traditional framework algorithm, the imaging performance is effectively improved, and the FPGA-based accelerator outperforms the CPU software implementation by 935 $\times $ .
- Open Access
Feature Pyramid U-Net with Attention for Semantic Segmentation of Forward-Looking Sonar Images
Forward-looking sonar is a technique widely used for underwater detection. However, most sonar images have underwater noise and low resolution due to their acoustic properties. In recent years, the semantic segmentation model U-Net has shown excellent segmentation performance, and it has great potential in forward-looking sonar image segmentation. However, forward-looking sonar images are affected by noise, which prevents the existing U-Net model from segmenting small objects effectively. Therefore, this study presents a forward-looking sonar semantic segmentation model called Feature Pyramid U-Net with Attention (FPUA). This model uses residual blocks to improve the training depth of the network. To improve the segmentation accuracy of the network for small objects, a feature pyramid module combined with an attention structure is introduced. This improves the model’s ability to learn deep semantic and shallow detail information. First, the proposed model is compared against other deep learning models and on two datasets, of which one was collected in a tank environment and the other was collected in a real marine environment. To further test the validity of the model, a real forward-looking sonar system was devised and employed in the lake trials. The results show that the proposed model performs better than the other models for small-object and few-sample classes and that it is competitive in semantic segmentation of forward-looking sonar images.
Open AccessFeature Pyramid U-Net with Attention for Semantic Segmentation of Forward-Looking Sonar Images
Forward-looking sonar is a technique widely used for underwater detection. However, most sonar images have underwater noise and low resolution due to their acoustic properties. In recent years, the semantic segmentation model U-Net has shown excellent segmentation performance, and it has great potential in forward-looking sonar image segmentation. However, forward-looking sonar images are affected by noise, which prevents the existing U-Net model from segmenting small objects effectively. Therefore, this study presents a forward-looking sonar semantic segmentation model called Feature Pyramid U-Net with Attention (FPUA). This model uses residual blocks to improve the training depth of the network. To improve the segmentation accuracy of the network for small objects, a feature pyramid module combined with an attention structure is introduced. This improves the model’s ability to learn deep semantic and shallow detail information. First, the proposed model is compared against other deep learning models and on two datasets, of which one was collected in a tank environment and the other was collected in a real marine environment. To further test the validity of the model, a real forward-looking sonar system was devised and employed in the lake trials. The results show that the proposed model performs better than the other models for small-object and few-sample classes and that it is competitive in semantic segmentation of forward-looking sonar images.