Xiaofei Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Xiaofei Wang has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
Pingyu Liu
Tuda Tong Intelligent Technology (Suzhou) Company, Ltd., Suzhou, China
2 shared publicationsRunxing Chen
Heilongjiang University
2 shared publicationsShuo Huang
Heilongjiang University
2 shared publicationsYong Liu
Heilongjiang University
2 shared publicationsYu Wang
Heilongjiang University
2 shared publications- Chuhan Qi1 shared publication
Latest Publications
- 31 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 23 (1545598X) , 1-5
Anomaly Detection Algorithm for Hyperspectral Images Based on Gabor Constraint Collaborative Representation
To address issues such as the insufficient utilization of spatial information in hyperspectral imagery (HSI), this paper proposes a novel hyperspectral anomaly detection algorithm based on Gabor constraint collaborative representation (GCCRD), building on the traditional collaborative representation detector (CRD) algorithm. First, spatial texture cues and spectral discrepancies between the test pixel and its surrounding circular window are jointly incorporated into a weighting model to enhance background representation. Second, a local consistency constraint is introduced to identify reliable background samples and suppress isolated noise by examining the spatial–spectral consistency among neighboring background pixels. Third, to avoid excessive suppression of anomalous targets, a saliency refinement strategy is adopted by combining the original saliency map with its Gaussian-smoothed counterpart. Experimental results on five real datasets demonstrate that the proposed GCCRD algorithm outperforms six state-of-the-art methods.
- 31 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 23 (1545598X) , 1-5
Anomaly Detection Algorithm for Hyperspectral Images Based on Gabor Constraint Collaborative Representation
To address issues such as the insufficient utilization of spatial information in hyperspectral imagery (HSI), this paper proposes a novel hyperspectral anomaly detection algorithm based on Gabor constraint collaborative representation (GCCRD), building on the traditional collaborative representation detector (CRD) algorithm. First, spatial texture cues and spectral discrepancies between the test pixel and its surrounding circular window are jointly incorporated into a weighting model to enhance background representation. Second, a local consistency constraint is introduced to identify reliable background samples and suppress isolated noise by examining the spatial–spectral consistency among neighboring background pixels. Third, to avoid excessive suppression of anomalous targets, a saliency refinement strategy is adopted by combining the original saliency map with its Gaussian-smoothed counterpart. Experimental results on five real datasets demonstrate that the proposed GCCRD algorithm outperforms six state-of-the-art methods.
- Open Access
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 19 (19391404) , 4163-4176
PyramidMamba: An Effective Hyperspectral Remote Sensing Image Target Detection Network
The lack of prior knowledge is a challenging issue in target detection tasks for hyperspectral remote sensing images. In this paper, we propose an effective network for object detection in hyperspectral remote sensing images. First, through spectral data augmentation methods, all surrounding pixels within a data block are encoded as the transformed spectral signature of the central pixel, thereby constructing a sufficient number of training sample pairs. Subsequently, a backbone network (PyramidMamba) was designed to establish long-term dependencies across the frequency domain and multi-scale dimensions using the Mamba residual module and pyramid wavelet transform module. A residual self-attention module is further developed, integrating self-attention with convolutional operations to enhance feature extraction while improving the network's depth and stability. A backbone network was employed to extract representative vectors from augmented sample pairs, which were then optimized through a spectral contrast head to enhance the distinction between target and background features. Experimental results demonstrate that compared to mainstream algorithms, the proposed algorithm achieves higher detection accuracy and computational efficiency. It successfully learns deep nonlinear feature representations with stronger discriminative power, enabling effective separation of targets from background and delivering state-of-the-art performance.
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 19 (19391404) , 4163-4176
Open AccessPyramidMamba: An Effective Hyperspectral Remote Sensing Image Target Detection Network
The lack of prior knowledge is a challenging issue in target detection tasks for hyperspectral remote sensing images. In this paper, we propose an effective network for object detection in hyperspectral remote sensing images. First, through spectral data augmentation methods, all surrounding pixels within a data block are encoded as the transformed spectral signature of the central pixel, thereby constructing a sufficient number of training sample pairs. Subsequently, a backbone network (PyramidMamba) was designed to establish long-term dependencies across the frequency domain and multi-scale dimensions using the Mamba residual module and pyramid wavelet transform module. A residual self-attention module is further developed, integrating self-attention with convolutional operations to enhance feature extraction while improving the network's depth and stability. A backbone network was employed to extract representative vectors from augmented sample pairs, which were then optimized through a spectral contrast head to enhance the distinction between target and background features. Experimental results demonstrate that compared to mainstream algorithms, the proposed algorithm achieves higher detection accuracy and computational efficiency. It successfully learns deep nonlinear feature representations with stronger discriminative power, enabling effective separation of targets from background and delivering state-of-the-art performance.
- Open Access
- 17 March 2025
- Published by Auricle Global Society of Education and Research in Computer Fraud and Security
- p. 120-130
Residual Network with Multi-scale Information Fusion for Apple Fungal Infection Classification
Black rot, Cedar rust, and Scab are three common apple diseases caused by fungal infections. The infected areas not only impair the function of the current parts but also spread to healthy regions of the plant, necessitating early detection in production. Addressing the challenge of subtle lesion features, we have designed a deep learning model that integrates multi-scale information fusion by combining wavelet transform with a residual network, and introducing a multi-channel Collaborative Attention mechanism. In comparisons with similar models, our proposed model achieved the best results, with a classification accuracy of 98.1% for various lesions. It demonstrated excellent stability in data detection imaging under various conditions. This model provides a new perspective for apple lesion detection and may potentially be analogously applied to other types of detection in the future.
- 17 March 2025
- Published by Auricle Global Society of Education and Research in Computer Fraud and Security
- p. 120-130
Open AccessResidual Network with Multi-scale Information Fusion for Apple Fungal Infection Classification
Black rot, Cedar rust, and Scab are three common apple diseases caused by fungal infections. The infected areas not only impair the function of the current parts but also spread to healthy regions of the plant, necessitating early detection in production. Addressing the challenge of subtle lesion features, we have designed a deep learning model that integrates multi-scale information fusion by combining wavelet transform with a residual network, and introducing a multi-channel Collaborative Attention mechanism. In comparisons with similar models, our proposed model achieved the best results, with a classification accuracy of 98.1% for various lesions. It demonstrated excellent stability in data detection imaging under various conditions. This model provides a new perspective for apple lesion detection and may potentially be analogously applied to other types of detection in the future.
- 29 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 21 (1545598X) , 1-5
Spatial Features Acquisition for Airplane Target Detection in Hyperspectral Remote Sensing Using Virtual RGB Images
Hyperspectral target detection (HTD) stands as a pivotal research area within hyperspectral image (HSI) processing. Airplane are challenging to detect as small targets in HSI. Problems with limited target samples and insufficient characterization capacity. Therefore in this letter, we propose a HTD algorithm leveraging virtual RGB images for spatial features acquisition. Initially, we train the semantic segmentation network U-Net v2 using the generated Aerial images for semantic segmentation (AAISS) datasets and employ HSI to synthesize virtual RGB images to capture high-level spatial features in the data. Subsequently, we fuse spatial and spectral features, culminating in pixel-level detection outcomes via the Random Forest (RF) algorithm. Experiments on real HSI datasets show that the proposed algorithm exhibits excellent target detection performance compared to the comparison algorithms.
- 29 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 21 (1545598X) , 1-5
Spatial Features Acquisition for Airplane Target Detection in Hyperspectral Remote Sensing Using Virtual RGB Images
Hyperspectral target detection (HTD) stands as a pivotal research area within hyperspectral image (HSI) processing. Airplane are challenging to detect as small targets in HSI. Problems with limited target samples and insufficient characterization capacity. Therefore in this letter, we propose a HTD algorithm leveraging virtual RGB images for spatial features acquisition. Initially, we train the semantic segmentation network U-Net v2 using the generated Aerial images for semantic segmentation (AAISS) datasets and employ HSI to synthesize virtual RGB images to capture high-level spatial features in the data. Subsequently, we fuse spatial and spectral features, culminating in pixel-level detection outcomes via the Random Forest (RF) algorithm. Experiments on real HSI datasets show that the proposed algorithm exhibits excellent target detection performance compared to the comparison algorithms.
- 3 May 2024
- Published by Taylor & Francis in International Journal of Remote Sensing
- Vol. 45 (10) , 3231-3247
DEDU-Net: Dual-Encoder-Decoder-U-Net for road extraction from high-resolution remote sensing images
- 3 May 2024
- Published by Taylor & Francis in International Journal of Remote Sensing
- Vol. 45 (10) , 3231-3247
DEDU-Net: Dual-Encoder-Decoder-U-Net for road extraction from high-resolution remote sensing images
- 27 February 2023
- Published by SPIE-Intl Soc Optical Eng
- Vol. 12585, 213-229
Research on method of avoiding phase unwrapping error in 3D measurement of gray code
In the field of industrial automation, accurate acquisition of object three-dimensional information is beneficial to better product design. Structured light three-dimensional measurement technique has been widely used in industrial inspection and reverse engineering because of its advantages of non-contact and high measurement accuracy. The structured light three-dimensional measurement technique combined with gray code and phase shift method has high efficiency and strong robustness. However, in the process of extract the phase map of the measured objects, there will be misalignment between the wrapped phase and the fringe order. This will lead to the jump error of wrapping phase edge in phase unwrapping process, which will reduce the measurement accuracy. To solve this problem, the fast tripartite phase unwrapping method is proposed in this paper, which can effectively eliminate the phase unwrapping error and significantly improve the computational efficiency. Firstly, the mask image is obtained by threshold segmentation of full-bright images using OTSU method. Subsequently, phase unwrapping is carried out under the guidance of gray code coding sequence, and the absolute phase value of the target is obtained. In the process of phase unwrapping, the mask image is used as the guide to improve the phase decoding efficiency, and tripartite phase unwrapping method is used to eliminate the phase unwrapping error. Experimental results show that this algorithm eliminates the edge jump error and improves the efficiency of normal gray code phase unwrapping method.
- 27 February 2023
- Published by SPIE-Intl Soc Optical Eng
- Vol. 12585, 213-229
Research on method of avoiding phase unwrapping error in 3D measurement of gray code
In the field of industrial automation, accurate acquisition of object three-dimensional information is beneficial to better product design. Structured light three-dimensional measurement technique has been widely used in industrial inspection and reverse engineering because of its advantages of non-contact and high measurement accuracy. The structured light three-dimensional measurement technique combined with gray code and phase shift method has high efficiency and strong robustness. However, in the process of extract the phase map of the measured objects, there will be misalignment between the wrapped phase and the fringe order. This will lead to the jump error of wrapping phase edge in phase unwrapping process, which will reduce the measurement accuracy. To solve this problem, the fast tripartite phase unwrapping method is proposed in this paper, which can effectively eliminate the phase unwrapping error and significantly improve the computational efficiency. Firstly, the mask image is obtained by threshold segmentation of full-bright images using OTSU method. Subsequently, phase unwrapping is carried out under the guidance of gray code coding sequence, and the absolute phase value of the target is obtained. In the process of phase unwrapping, the mask image is used as the guide to improve the phase decoding efficiency, and tripartite phase unwrapping method is used to eliminate the phase unwrapping error. Experimental results show that this algorithm eliminates the edge jump error and improves the efficiency of normal gray code phase unwrapping method.
- 16 February 2023
- Published by Taylor & Francis in International Journal of Remote Sensing
- Vol. 44 (4) , 1238-1253
Dynamic programming-based band selection method for hyperspectral unmixing
Band selection (BS) is a method for optimizing feature selection, which aims to of reduce the computational complexity of processing hyperspectral image (HSI). However, there are many BS methods applied to image classification, target detection, and anomaly detection. Furthermore, the existing BS methods ignore the spatial structure of HSI. To solve the above problems, we proposed a dynamic programming-based BS method for hyperspectral unmixing. In this paper, we use the convex geometric structure of HSI in band space to project it into the subspace to obtain depth spectral features, then construct a dynamic programming model to select representative bands. To verify the effectiveness of the proposed method, experiments are conducted on three widely used datasets, and compared with three popular BS methods. The experimental results show that the proposed method has satisfactory performance in different evaluation indexes (including signal to reconstruction error (SRE), root mean square error (RMSE)) and three quantitative evaluations (average information entropy (AIE), average correlation coefficient (ACC) and average relative entropy (ARE)).
- 16 February 2023
- Published by Taylor & Francis in International Journal of Remote Sensing
- Vol. 44 (4) , 1238-1253
Dynamic programming-based band selection method for hyperspectral unmixing
Band selection (BS) is a method for optimizing feature selection, which aims to of reduce the computational complexity of processing hyperspectral image (HSI). However, there are many BS methods applied to image classification, target detection, and anomaly detection. Furthermore, the existing BS methods ignore the spatial structure of HSI. To solve the above problems, we proposed a dynamic programming-based BS method for hyperspectral unmixing. In this paper, we use the convex geometric structure of HSI in band space to project it into the subspace to obtain depth spectral features, then construct a dynamic programming model to select representative bands. To verify the effectiveness of the proposed method, experiments are conducted on three widely used datasets, and compared with three popular BS methods. The experimental results show that the proposed method has satisfactory performance in different evaluation indexes (including signal to reconstruction error (SRE), root mean square error (RMSE)) and three quantitative evaluations (average information entropy (AIE), average correlation coefficient (ACC) and average relative entropy (ARE)).
- Open Access
- 25 August 2020
- Published by Wiley in Mathematical Problems in Engineering
- Vol. 2020, 1-12
An Overview on Linear Unmixing of Hyperspectral Data
Hyperspectral remote sensing technology has a strong capability for ground object detection due to the low spatial resolution of hyperspectral imaging spectrometers. A single pixel that leads to a hyperspectral remote sensing image usually contains more than one feature coverage type, resulting in a mixed pixel. The existence of a mixed pixel affects the accuracy of the ground object identification and classification and hinders the application and development of hyperspectral technology. For the problem of unmixing of mixed pixels in hyperspectral images (HSIs), the linear mixing model can model the mixed pixels well. Through the collation of nearly five years of the literature, this paper introduces the development status and problems of linear unmixing models from four aspects: geometric method, nonnegative matrix factorization (NMF), Bayesian method, and sparse unmixing.
- 25 August 2020
- Published by Wiley in Mathematical Problems in Engineering
- Vol. 2020, 1-12
Open AccessAn Overview on Linear Unmixing of Hyperspectral Data
Hyperspectral remote sensing technology has a strong capability for ground object detection due to the low spatial resolution of hyperspectral imaging spectrometers. A single pixel that leads to a hyperspectral remote sensing image usually contains more than one feature coverage type, resulting in a mixed pixel. The existence of a mixed pixel affects the accuracy of the ground object identification and classification and hinders the application and development of hyperspectral technology. For the problem of unmixing of mixed pixels in hyperspectral images (HSIs), the linear mixing model can model the mixed pixels well. Through the collation of nearly five years of the literature, this paper introduces the development status and problems of linear unmixing models from four aspects: geometric method, nonnegative matrix factorization (NMF), Bayesian method, and sparse unmixing.
- Open Access
- 28 July 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 8, 138067-138076
GF-2 Panchromatic and Multispectral Remote Sensing Image Registration Algorithm
Image registration is an important step in remote sensing image preprocessing. The accuracy, efficiency, and automatic degree of image registration will directly affect the results of subsequent image processing and analysis. Due to the slow registration speed and low registration accuracy of SIFT algorithm for GF-2 panchromatic and multispectral remote sensing images, SIFT algorithm is improved in this article to improve the efficiency of image registration algorithm. In this article, the strategy of information entropy meshes is introduced, and feature extraction is carried out only for the regions with large information entropy, so that the running time of feature extraction can be reduced. By introducing Canny operator to eliminate the unstable edge response points, the number of feature points can be further reduced, thus the computation amount of the algorithm can be also reduced. Due to the poor registration effect of SIFT algorithm for remote sensing images, a new gradient calculation method and feature description method are used in this article to enhance the robustness of feature descriptors. In the feature matching stage, this article proposes a two-layer matching strategy, which uses mutual Euclidean distance for initial matching, then uses FSC algorithm for fine matching, and finally realizes the registration of GF-2 panchromatic and multispectral remote sensing images. Experimental results show that the proposed algorithm can obtain more correct matching point pairs when the number of extracted feature points is small. Moreover, the registration accuracy and speed are both higher than the comparison algorithm. This article uses the remote sensing images of urban scenes dominated by plains and mountainous scenes with small terrain fluctuation respectively for experiments. The algorithm in this article can achieve registration accuracy better than 0.5 pixels, and in terms of time consumption, it is only about 70% of SIFT algorithm, and the computational efficiency is better than SIFT algorithm, which can meet the actual GF-2 Pan and MSI registration tasks.
- 28 July 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 8, 138067-138076
Open AccessGF-2 Panchromatic and Multispectral Remote Sensing Image Registration Algorithm
Image registration is an important step in remote sensing image preprocessing. The accuracy, efficiency, and automatic degree of image registration will directly affect the results of subsequent image processing and analysis. Due to the slow registration speed and low registration accuracy of SIFT algorithm for GF-2 panchromatic and multispectral remote sensing images, SIFT algorithm is improved in this article to improve the efficiency of image registration algorithm. In this article, the strategy of information entropy meshes is introduced, and feature extraction is carried out only for the regions with large information entropy, so that the running time of feature extraction can be reduced. By introducing Canny operator to eliminate the unstable edge response points, the number of feature points can be further reduced, thus the computation amount of the algorithm can be also reduced. Due to the poor registration effect of SIFT algorithm for remote sensing images, a new gradient calculation method and feature description method are used in this article to enhance the robustness of feature descriptors. In the feature matching stage, this article proposes a two-layer matching strategy, which uses mutual Euclidean distance for initial matching, then uses FSC algorithm for fine matching, and finally realizes the registration of GF-2 panchromatic and multispectral remote sensing images. Experimental results show that the proposed algorithm can obtain more correct matching point pairs when the number of extracted feature points is small. Moreover, the registration accuracy and speed are both higher than the comparison algorithm. This article uses the remote sensing images of urban scenes dominated by plains and mountainous scenes with small terrain fluctuation respectively for experiments. The algorithm in this article can achieve registration accuracy better than 0.5 pixels, and in terms of time consumption, it is only about 70% of SIFT algorithm, and the computational efficiency is better than SIFT algorithm, which can meet the actual GF-2 Pan and MSI registration tasks.