A5050009113- Remote-Sensing Image Classification
- Remote Sensing and Land Use
- Advanced Image Fusion Techniques
- Remote Sensing in Agriculture
- Land Use and Ecosystem Services
- Remote Sensing and LiDAR Applications
- Image and Signal Denoising Methods
- Domain Adaptation and Few-Shot Learning
- Sparse and Compressive Sensing Techniques
- Flood Risk Assessment and Management
- Coastal wetland ecosystem dynamics
- Image Retrieval and Classification Techniques
- Spectroscopy and Chemometric Analyses
- Machine Learning and ELM
- Advanced Image and Video Retrieval Techniques
- Urban Heat Island Mitigation
- 3D Shape Modeling and Analysis
- Face and Expression Recognition
- Image Enhancement Techniques
- Coastal and Marine Dynamics
- Precipitation Measurement and Analysis
- Environmental Changes in China
- Human Pose and Action Recognition
- Endometriosis Research and Treatment
- Landslides and related hazards
Ningbo University
2016-2025
University of Nottingham Ningbo China
2021-2025
Shanghai Technical Institute of Electronics & Information
2025
Southeast University
2023-2024
Ragon Institute of MGH, MIT and Harvard
2020-2024
Qingdao Binhai University
2024
Qingdao University of Science and Technology
2024
China Agricultural University
2024
Southwest Forestry University
2023-2024
Shanghai Municipal Center For Disease Control Prevention
2024
Many problems in computer vision require dealing with sparse, unordered data the form of point clouds. Permutation-equivariant networks have become a popular solution - they operate on individual points simple perceptrons and extract contextual information global pooling. This can be achieved normalization feature maps, operation that is unaffected by order. In this paper, we propose Attentive Context Normalization (ACN), yet effective technique to build permutation-equivariant robust...
Spatial-temporal-spectral fusion (STSF) of remote sensing imagery can produce data with the highest spatial and spectral resolution, only as well fine temporal by integrating images complementary information in both domains. Accuracy variation is an important guarantee for achieving fidelity STSF. However, current STSF methods estimate utilizing between observed multispectral image (MSI) relationship MSI hyperspectral (HSI), which difficult to obtain accurate variation. To address this...
The monitoring of coastal wetlands is great importance to the protection marine and terrestrial ecosystems. However, due complex environment, severe vegetation mixture, difficulty access, it impossible accurately classify identify their species with traditional classifiers. Despite integration multisource remote sensing data for performance enhancement, there are still challenges acquiring exploiting complementary merits from data. In this article, depthwise feature interaction network...
With the fast development of remote sensing platforms and sensors technology, change detection with heterogeneous images (Hete-CD) has become an attractive topic in recent years plays a vital role land cover for responding to natural disaster emergencies when homogeneous are unavailable. Although Hete-CD been developed about three decades, various related methods have applied successfully practice, systematic comprehensive review current achievements regarding remains lacking. Therefore,...
Land cover change detection (LCCD) using remote sensing images (RSIs) plays an important role in natural disaster evaluation, forest deformation monitoring, and wildfire destruction detection. However, bitemporal are usually acquired at different atmospheric conditions, such as sun height soil moisture, which cause pseudo noise into the map. Changed areas on ground also generally have various shapes sizes, consequently making utilization of spatial contextual information a challenging task....
Recent studies have shown that deep domain adaptation (DA) techniques good performance on cross-domain hyperspectral image (HSI) classification problems. However, most existing HSI DA approaches directly use networks to extract features from the data, which ignores detailed information of in spectral and spatial dimensions. To effectively exploit spectral–spatial joint for HSIs, we propose a two-branch attention adversarial (TAADA) network this article. In TAADA network, feature extraction...
Deep learning (DL) based methods represented by convolutional neural networks (CNNs) are widely used in hyperspectral image classification (HSIC). Some of these have strong ability to extract local information, but the extraction long-range features is slightly inefficient, while others just opposite. For example, limited receptive fields, CNN difficult capture contextual spectral-spatial from a relationship. Besides, success DL-based greatly attributed numerous labeled samples, whose...
Traditional remote sensing (RS) image classification methods heavily rely on labeled samples for model training. When are unavailable or have different distributions from that of the to be classified, may fail. The cross-domain cross-scene is developed this case where an existing training and unknown scenes domains classification. distribution inconsistency problem caused by differences in acquisition environment conditions, scene, time or/and changing sensors. To cope with problem, many...
The accurate extraction and mapping of floating raft aquaculture (FRA) is significant to the scientific management sustainable development coastal zones. However, current relevant methods rely on large sample size complex classifiers, which have poor generalization ability thus are not suitable for large-scale application. To address these issues, this study proposes a new hyperspectral index based remote sensing images, namely (HSI-FRA). Based analysis spectral information, HSI-FRA utilizes...
Spectral-spatial features are important for ground target identification and classification with High Spatial Resolution Remotely Sensed (HSRRS) Imagery. In this paper, two novel features, named the Gaussian-Weighting Spectral (GWS) feature Area Shape Index (ASI) feature, proposed to complement deficiency of basic image land cover HSRRS imagery. The GWS is an adaptive region-based that aims improve spectral homogeneity a local area surrounding pixel. Additionally, it well known inadequate...
Labeled samples are important in achieving land cover change detection (LCCD) tasks via deep learning techniques with remote sensing images. However, labeling for bitemporal images is labor-intensive and time-consuming. Moreover, manually between requires professional knowledge practitioners. To address this problem article, an iterative training sample augmentation (ITSA) strategy to couple a neural network improving LCCD performance proposed here. In the ITSA, we start by measuring...
Mangroves play a significant role in carbon sequestration and storage. Mapping mangrove species monitoring their conditions have been crucial issue for achieving sustainable development goals. Currently combing multidimensional optical SAR images with machine learning become an important approach classification, but there are still some challenges feature selection hyperparameter optimizations. In this study, we proposed novel classification framework by multi-scale variable algorithm (MUVR)...
Recently, prototypical network based few-shot learning (FSL) has been introduced for small-sample hyperspectral image (HSI) classification and shown good performance. However, existing prototypical-based FSL methods have two problems: prototype instability domain shift between training testing datasets. To solve these problems, we propose a refined contrastive (RPCL-FSL) in this paper, which incorporates supervised into an end-to-end to perform HSI classification. stabilize refine the...
An improved sparse subspace clustering (ISSC) method is proposed to select an appropriate band subset for hyperspectral imagery (HSI) classification. The ISSC assumes that vectors are sampled from a union of low-dimensional orthogonal subspaces and each can be sparsely represented as linear or affine combination other bands within its subspace. First, the represents with coefficient by solving L2-norm optimization problem using least square regression (LSR) algorithm. block diagonal...
A low-rank and sparse matrix decomposition (LRaSMD) detector has been proposed to detect anomalies in hyperspectral imagery (HSI). The assumes background images are while gross errors that sparsely distributed throughout the image scene. By solving a constrained convex optimization problem, LRaSMD separates from background. This protects model corruption. An anomaly value for each pixel is calculated using Euclidean distance, determined by thresholding value. Four groups of experiments on...
A fast and robust principal component analysis on Laplacian graph (FRPCALG) method is proposed to select bands of hyperspectral imagery (HSI). The FRPCALG assumes that a clean band matrix lies in unified manifold subspace with low-rank clustering properties, whereas sparse noise does not lie the same subspace. It estimates lowrank approximation original HSI while uncovering structure all bands. Specifically, structured random projection adopted reduce high spatial dimensionality data for...
This article introduces the design and imaging principles of Advanced Hyperspectral Imager (AHSI) aboard China's GaoFen-5 satellite. The AHSI is a visible nearinfrared (VNIR)/short-wave infrared (SWIR) HSI. It first spaceborne hyperspectral sensor that utilizes both convex-grating spectrophotometry an improved three-concentric-mirror (Offner) configuration. has 330 spectral bands, 60-km swath width, 30-m spatial resolution. Various tests have been designed to evaluate its performance,...