Deblurring in remote sensing images is a challenging task due to the long-range imaging capabilities of remote sensing sensors, which often results in image blur. Factors contributing to image blur include atmospheric disturbances during long-range imaging or the orbital motion of remote sensing platforms. The existing methods remove blur in remote sensing images using the traditional attention mechanism, which focuses on a limited number of features. However, they often overlook the features among neighboring positions in blurry areas, and these areas contain more relevant features. Leveraging these features can effectively assist in restoring the complex object textures of remote sensing blurry images. To achieve this, we propose a novel implicit neural attention mechanism for assembling more relevant features implied by surrounding coordinates. Specifically, we use the features and their corresponding coordinates to learn the enhanced feature representation with more relevant features, and this representation can be used to derive the deblurred images. Extensive experiments demonstrate that our proposed method, INA-RSDeblur, outperforms the state-of-the-art deblurring methods in remote sensing blurry images.

Existing CNN-based and Transformer-based methods have demonstrated remarkable performance in low-level visual tasks, including image deblurring. These methods generally capture spatial features only in a single way, such as by stacking blocks of CNNs and Transformers, resulting in inadequate utilization of spatial context. To address this issue, we propose a new feature aggregation scheme for image deblurring, named Omni-Deblurring. The core of our omni-deblurring is the omni-range context block, which enables explicitly aggregating the local-range, regional-range, and global-range features in a compact manner. With this design, it can bring a wider receptive field for modeling the contextual features. Extensive experiments on synthetic and real-world blurry datasets demonstrate the effectiveness of our proposed method in both quantitative and qualitative evaluations. Furthermore, the quality of our deblurring model is evaluated in the task of object detection, and the mean Average Precision (mAP) metric increases by 10% across all classes compared with other deblurring models. Code is available athttps://github.com/yaowli468/Omni-Deblurring.

Graph convolution neural networks (GCNNs) have shown great promise in handling point cloud due to their adeptness at extracting topological information. However, a significant limitation of current GNN approaches is the substantial time required for constructing and learning the graph structures inherent in the point cloud. To address this challenge, we present a novel graph convolution-based network designed for the efficient processing of 3D point cloud. Based on dynamic GCNN (DGCNN) and PointNet++, our network integrates graph convolution with multi-scale point and edge features for enhanced neighborhood geometric information extraction. We introduce an innovative lightweight edge convolution technique that notably reduces computational burden while optimizing processing efficiency. In addition, the network also employs a downsampling strategy within a GCNN framework and introduces a directed graph-based geometric feature extractor to robustly capture local geometries. Empirically, our model demonstrates state-of-the-art performance across three benchmark datasets, markedly improving both accuracy and processing speed in point cloud analysis. Notably, our model achieves a groundbreaking 90% overall accuracy on the ScanObjectNN dataset.

Point-based methods have made significant progress, but improving their scalability in large-scale 3D scenes is still a challenging problem. In this paper, we delve into the point-based method and develop a simpler, faster, stronger variant model, dubbed as LinNet. In particular, we first propose the disassembled set abstraction (DSA) module, which is more effective than the previous version of set abstraction. It achieves more efficient local aggregation by leveraging spatial anisotropy and channel anisotropy separately. Additionally, by mapping 3D point clouds onto 1D space-filling curves, we enable parallelization of downsampling and neighborhood queries on GPUs with linear complexity. LinNet, as a purely point-based method, outperforms most previous methods in both indoor and outdoor scenes without any extra attention, and sparse convolution but merely relying on a simple MLP. It achieves the mIoU of 73.7\%, 81.4\%, and 69.1\% on the S3DIS Area5, NuScenes, and SemanticKITTI validation benchmarks, respectively, while speeding up almost 10x times over PointNeXt. Our work further reveals both the efficacy and efficiency potential of the vanilla point-based models in large-scale representation learning. Our code will be available upon publication.

Earth’s surface is covered by thin-cloud throughout the year, which greatly limits the application of remote sensing (RS) images obtained at a high cost. Currently, deep-learning technology has received widespread attention in the field of thin-cloud removal from RS images. The parameter training process of the deep network in this letter is divided into two stages according to the frequency domain feature of training samples. In the first stage, the loss between the thin-cloud removal image and the corresponding clear image is minimized by the full-frequency domain training. In the second stage, the parameters of the first stage are frozen, and the high-frequency features of the thin-cloud image are integrated into the Decoder module for network training. Furthermore, the corresponding attention module of the dual-tree wavelet is designed. For better training in the second stage, the high-frequency loss function based on three-level wavelet transform is designed. Compared with the highest average values of peak signal to noise ratio (PSNR) and structural similarity (SSIM) of traditional methods, FDT-Net achieves an improvement by up to 33.4108% and 7.5184%, respectively. In quantitative experiments, compared with the highest average values of PSNR and SSIM among deep-learning methods, FDT-Net has an improvement by 2.1505% and 0.9123%, respectively. Therefore, the proposed deep-learning network FDT-Net can significantly extend the temporal and spatial range of RS image application under thin-cloud weather conditions. The code of model is available at https://github.com/chonghaozhan/FDT-Net.

Different from the ground image with uniform haze, the haze in remote sensing (RS) image has the characteristics of irregular shape and uneven concentration in hazy weather. It brings a great challenge to the application of RS image data in advanced image processing tasks. A novel dehazing network for non-uniform hazy remote sensing image, named as KFA-Net, is proposed to solve the aforementioned issues. The designed asymmetric size feature cascade (ASFC), k-means pixel attention (KPA) and FFT channel attention (FCA) in KFA-Net all show excellent effects. Compared with symmetrically linked typical Unet, ASFC can more easily extract shallow features for feature reconstruction. Furthermore, different from the commonly used pixel attention that compresses feature maps directly, KPA introduces k-means clustering algorithm in machine learning into the attention mechanism, which facilitates network training to focus on the thick hazy region. Compared to typical squeeze-and-excitation block, FCA uses the low-frequency region feature of spectrogram to obtain the attention weight coefficient in the frequency domain, making network training pay more attention to the feature of image low-frequency region. Extensive comparison experiments verify that the proposed KFA-Net has the great superiority. PSNR/SSIM of KFA-Net are 31.0952% and 6.6401% higher than DCP with the highest citation in traditional dehazing methods, respectively. PSNR/SSIM of KFA-Net are 2.2049% and 0.4966% higher than the recently proposed 4KDehazing with the best performance among all comparison dehazing methods, respectively. The KFA-Net proposed in this research can greatly enhance the temporal and spatial scope of RS image application in hazy weather conditions.