Due to rich spectral information, hyperspectral images (HSIs) have been widely used in various fields. However, limited by imaging systems, the low spatial resolution of HSIs has become an important problem. In this article, for enhancing resolution, Interactformer is proposed interact with global and local features extracted Transformer 3D convolutional neural network (CNN) branches. Within branch, a separable self-attention module linear complexity designed solve problem that traditional...

Pan-sharpening is an effective method to obtain high-resolution multispectral images by fusing panchromatic (PAN) with fine spatial structure and low-resolution rich spectral information. In this article, a multiscale pan-sharpening based on dynamic convolutional neural network proposed. The filters in convolution are generated dynamically locally the filter generation which different from standard strengthens adaptivity of network. adaptively changed according input images. proposed...

Coastal wetland monitoring plays an important role in the protection and restoration of ecosystems this world. UAV-hyperspectral imaging, as emerging technique for Earth observation space exploration, provides huge potential ability to identify different species. In work, a multilayer global spectral–spatial attention network (MGSSAN) is proposed mapping coastal wetlands, which mainly consists two major steps. First, two-branch convolutional neural (CNN) framework with residual connection...

Crafting an edge-based real-time object detector for unmanned aerial vehicle (UAV) images is challenging because of the limited computational resources and small size detected objects. Existing lightweight detectors often prioritize speed over detecting extremely targets. To better balance this trade-off, paper proposes efficient low-complexity edge computing platforms deployed on UAVs, termed EUAVDet (Edge-based UAV Object Detector). Specifically, feature downsampling module a novel...

Automatic ultrasound image segmentation improves the efficiency of clinical diagnosis and decreases workload doctors. Many methods only focus on capturing local details global dependencies, whereas ignoring large-scale context information. However, it is essential to extract features for large targets in images. To enhance capability feature extraction model with various sizes improve performance, we propose an effective multilevel network (SLG-Net) which can from small details,...

In this paper we introduce the nonsubsampled shear let transform for multi-focus image fusion. proposed method, source images are decomposed by firstly. Then decomposition coefficients merged according to given fusion rule. Finally fused is reconstructed inverse transform. The experimental results over five pairs of registered and one pair mis-registered demonstrate superiority method.

The spectral resolution of hyperspectral images (HSIs) is very high. Nevertheless, their spatial low due to various hardware limitations. Therefore, it important study HSI super improve resolution. In this article, for single-image resolution, we propose a multiscale feature fusion and aggregation network with 3-D convolution (MFFA-3D) by cascading the MFFA-3D block. block includes group part part. part, novel method proposed. Group module two-step are proposed in order prevent distortion,...

Hyperspectral images (HSIs) contain rich spectral information and have great application value. However, due to various hardware limitations, the spatial resolution of HSIs acquired by sensor is low. HSI super-resolution (SR) attracts much attention improve quality. In this letter, a single SR method based on network fusion proposed. Our includes part part. part, we construct 3-D multiscale mixed networks (3-D-MSMANs) cascading block (3-D-MSMAB) restore high-resolution HSIs. 3-D-MSMAB...

Pansharpening is a spatial-spectral fusion technique that fuses low-resolution multispectral (MS) images with high-resolution panchromatic (PAN) to get which are rich in spectral and spatial information. Some pansharpening methods based on dynamic convolution were proposed improve the adaptivity generalization of network. However, these either only focus local small regions or generate filters complex Besides, existing convolve MS PAN image, resulting extracted details features inadequate....

Pansharpening is an effective technology to obtain high resolution multispectral (HRMS) images by fusing low (LRMS) and panchromatic (PAN) images. With the rapid development of deep learning, some pansharpening methods based on learning have been proposed. Although fused are greatly improved, there still areas for improvement. For example, spectral preservation not good enough details rich enough. To address above problems, a two-stage method convolutional neural network (CNN) In first...

Deep convolutional neural networks (CNNs) have made great progress in the super-resolution (SR) of hyperspectral images (HSIs). However, most methods utilize convolution to explore local features, and global features are ignored. It is expected that combining non-local mechanism with CNN will improve performance HSI SR. This paper presents a multi-level progressive SR network. The dense block (DNLB) constructed combine which used reconstruct at each level. Due high dimension HSI, original...

In this paper, we propose a novel feature vector clustering method for unsupervised change detection in multitemporal satellite images. A each pixel is extracted using the compressed sparse representation of difference image which obtained by comparing pair co-registered images acquired at different times on same area. The achieved taking two stages: sampling and representation. first employed order to reduce dimensionality vectors. Then, applied extract meaningful information combat noise...

This paper presents a novel method based on the developed multiscale dual bilateral filter to fuse high spatial resolution panchromatic image and spectral multispectral image. Compared with traditional multi-resolution methods, process of detail extraction considers characteristics simultaneously. The low is resampled same size sharpened through injecting extracted details. proposed fusion tested over QuickBird IKONOS images compared three popular methods. experimental results demonstrate...

Many hyperspectral image (HSI) super-resolution (SR) methods have been proposed and achieved good results; however, they do not sufficiently preserve the spectral information. It is beneficial to utilize correlation. In addition, most works super-resolve images using high computation complexity. To solve above problems, a novel method based on channel multilayer perceptron (CMLP) presented in this article, which aims obtain better performance while reducing computational cost. extract...

Image super resolution is a challenging highly ill-posed inverse problem. In this paper, we proposed texture constrained sparse representation for single image resolution. Firstly, the low observed segmented into different regions. Through preprepared databases, regions are classified categories using designed classifier. Then, high segments reconstructed by with relevant dictionaries. Integrating all segments, result obtained. The method compared and some existing methods. experimental...

Sparse representation using the over-complete dictionary makes that decomposition coefficients are more sparse, and can reflect inherent characteristics structure of signals. A novel fusion method based on IHS transform sparse for multi-spectral image panchromatic is proposed in this paper. Firstly, applied to image. Then, intensity component fused through representation. used local patch instead whole image, because size usually large which not suitable Experimental results indicate has...

Titanium carbide–titanium diboride (TiC–TiB2) composite powders were synthesised through a carbothermal reduction method by using titanium dioxide, boric acid, and different carbon sources (namely, black, sucrose, glucose) as starting materials. The thermal decomposition behaviour of the precursors was studied thermogravimetry–differential analyser. Phase compositions morphologies products characterised X-ray diffractometer scanning electron microscope. When n(Ti):n(B):n(C) = 1.0:2.5:5.0,...