A Novel Hyperspectral Image Change Detection Framework Based on 3D-Wavelet Domain Active Convolutional Neural Network

Wang, Xianghai; Xing, Chengdi; Feng, Yining; Song, Ruoxi; Mu, Zhenhua

doi:10.1109/igarss47720.2021.9553050

Cited by 14 publications

(6 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Unfortunately, the 2D difference matrix is a symmetric matrix, so there is at least half of the invalid information for the detection model. Detection accuracy degradation caused by the ability of the detection model in terms of difference feature capture [11, 19]. Change feature capture techniques based on traditional methods (e.g., improved change vector analysis (ICVA) [20]) and deep learning (DL) (e.g., convolutional neural networks (CNN) [21]) methods are often seen in HSI‐CD tasks in recent years. The former aims to use hand‐crafted features for change identification of bitemporal HSIs, while the latter employs a data‐driven approach to detect changing pattern at the semantic‐level feature.…”

Section: Introductionmentioning

confidence: 99%

“…Over the past 10 years, the DL methods and their variants, such as the CNN [21,22], recurrent neural network (RNN) [23], autoencoder [9,24], graph convolutional network (GCN) [25], residual network (ResNet) [26], principal component analysis network (PCANet) [27], etc., have emerged in the HSI-CD fields. In particular, the CNN-based HSI-CD framework is favoured by researchers due to its universality and portability [28].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Spectral‐spatial sequence characteristics‐based convolutional transformer for hyperspectral change detection

Zhou

Shi

et al. 2023

CAAI Trans on Intel Tech

View full text Add to dashboard Cite

Recently, ground coverings change detection (CD) driven by bitemporal hyperspectral images (HSIs) has become a hot topic in the remote sensing community. There are two challenges in the HSI‐CD task: (1) attribute feature representation of pixel pairs and (2) feature extraction of attribute patterns of pixel pairs. To solve the above problems, a novel spectral‐spatial sequence characteristics‐based convolutional transformer (S3C‐CT) method is proposed for the HSI‐CD task. In the designed method, firstly, an eigenvalue extrema‐based band selection strategy is introduced to pick up spectral information with salient attribute patterns. Then, a 3D tensor with spectral‐spatial sequence characteristics is proposed to represent the attribute features of pixel pairs in the bitemporal HSIs. Next, a fusion framework of the convolutional neural network (CNN) and Transformer encoder (TE) is designed to extract high‐order sequence semantic features, taking into account both local context information and global sequence dependencies. Specifically, a spatial‐spectral attention mechanism is employed to prevent information reduction and enhance dimensional interactivity between the CNN and TE. Finally, the binary change map is determined according to the fully‐connected layer. Experimental results on real HSI datasets indicated that the proposed S3C‐CT method outperforms other well‐known and state‐of‐the‐art detection approaches in terms of detection performance.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Spectral‐spatial sequence characteristics‐based convolutional transformer for hyperspectral change detection

Zhou

Shi

et al. 2023

CAAI Trans on Intel Tech

View full text Add to dashboard Cite

show abstract

“…However, these methods cannot completely remove the influence of noise in the process of CD, so scholars have proposed some new methods for the generation of DI, such as stacked autoencoders (SAEs) [19] and deep pyramid feature learning networks (DPFL-Nets) [20]. Moreover, multiscale decomposition of DI is also a common method, such as principal component analysis (PCA) [21], nonsubsampled shearlet transform (NSST) [18], discrete 3D-wavelet transform (3-DWT) [22], and scale-invariant feature transform (SIFT) [23]. Although these methods can better overcome the influence of noise, the disadvantage is that the algorithm runs for a long time.…”

Section: Introductionmentioning

confidence: 99%

Unsupervised Change Detection in Wide-Field Video Images Under Low Illumination

Shi

Jia

Yang

et al. 2023

IEEE Trans. Circuits Syst. Video Technol.

View full text Add to dashboard Cite

show abstract

“…1) Firstly, multi-temporal remote sensing images are preprocessed to provide high-quality data input (correction, enhancement, and registration [21] of original data, etc.) 2) Secondly, feature extraction and selection (spectral, spatial, object and scene) can be carried out [22], [23]. In this process, feature fusion [24] at different levels can be carried out, which is related to the change detection algorithm model.…”

Section: Introductionmentioning

confidence: 99%

SSCFNet: A Spatial-Spectral Cross Fusion Network for Remote Sensing Change Detection

Wang

Liu

Jiao

et al. 2023

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Convolutional Neural Networks(CNNs) are datadriven methods that automatically extract the rich information embedded in remote sensing images. However, most current deep learning-based remote sensing image change detection methods prioritize high-level semantic features, while not enough attention is given to low-level semantic features, resulting in the loss of edges and details of the change region. To address this problem, this paper constructs a spatial-spectral cross fusion network SSCFNet, divided into three modules: a feature extractor network module, a combined enhancement module, and a semantic cross-fusion module. A new combined enhancement strategy is proposed to construct several semantic feature blocks in the combined enhancement module. Different convolution operations are applied to the newly constructed semantic feature blocks in the semantic cross fusion module, and the obtained semantic features at various levels are cross-fused. Experiments show that the proposed SSCFNet outperforms the other six state-of-the-art methods on four publicly available remote sensing image change detection datasets.

show abstract

A Novel Hyperspectral Image Change Detection Framework Based on 3D-Wavelet Domain Active Convolutional Neural Network

Cited by 14 publications

References 7 publications

Spectral‐spatial sequence characteristics‐based convolutional transformer for hyperspectral change detection

Spectral‐spatial sequence characteristics‐based convolutional transformer for hyperspectral change detection

Unsupervised Change Detection in Wide-Field Video Images Under Low Illumination

SSCFNet: A Spatial-Spectral Cross Fusion Network for Remote Sensing Change Detection

Contact Info

Product

Resources

About