Existing cloud detection methods often rely on deep neural networks, leading to excessive computational overhead. To address this, we propose a shallow convolutional neural network (CNN)-Transformer hybrid architecture that limits the maximum downsampling rate to 8×. This design preserves local details while effectively capturing global context through a lightweight Transformer branch. To enhance adaptability across diverse cloud scenes, we introduce two novel statistics-driven modules: statistics-adaptive convolution (SAC) and statistical mixing augmentation (SMA). SAC dynamically generates convolutional kernels based on input feature statistics, enabling adaptive feature extraction for varying cloud patterns. SMA improves model generalization by interpolating channel-wise statistics across training samples, increasing feature diversity. Experiments on four datasets show that the proposed method achieves state-of-the-art performance with 732K parameters and 1G multiply-accumulate operations. Code will be available at https://weix-liu.github.io/ for further research.

Domain adaptive object detection methods for remote sensing images typically rely on large-scale target domain data and multi-epoch offline adaption training. However, the wide variation in remote sensing conditions makes it difficult to gather sufficient data for every potential target domain, especially for unexpected domains. To address this challenge, we propose Learnable Online Style Adaptation (LOSA), a method that enables the source model to adapt effectively to new target domain styles with low test-time latency. Specifically, LOSA captures target domain styles using shallow feature channel statistics and predicts style shifts based on channel dependencies to recalibrate target features. Through a coarse-to-fine alignment loss between online target features and pre-computed source domain statistics, LOSA autonomously learns domain-specific style adaptation strategies. By adopting a dynamically optimized high learning rate, LOSA only requires a small number of samples for test-time training, making it suitable for real-time applications. Experimental results across various scenarios, including normal-to-corrupted, cross-band, and sim-to-real adaptation, demonstrate that the proposed method significantly improves cross-domain object detection performance. Moreover, our method can be well generalized to cross-domain image classification tasks.

In the field of simultaneous localization and mapping (SLAM), map-based localization has been widely used in autonomous driving, particularly for all-speed and all-road adaptive cruise, automatic parking, and other high-level functions. As a result, LiDAR sensors are frequently used in visual-based SLAM to improve the overall accuracy of ego-motion estimation and environment reconstruction. In this article, a novel tightly coupled monocular hybrid visual LiDAR SLAM (HVL-SLAM), which utilizes both visual and LiDAR measurements in tracking and mapping. First, the proposed method reduces the 3-D uncertainty of features by employing object segmentation and Delaunay triangulation. The motion between adjacent frames is then estimated using a hybrid tracking module that minimizes photometric and reprojection error. Finally, a joint optimization method for refining the pose is proposed, which incorporates visual and LiDAR measurements into optimization with dynamic weights, resulting in higher positioning accuracy and robustness. The experiments on the public KITTI odometry benchmark and real-world outdoor datasets demonstrate that HVL-SLAM outperforms state-of-the-art approaches in terms of pose estimation and mapping performance. The code is released to the community. Code available athttps://github.com/kinggreat24/hvl_slam.

Vehicle detection methods based on deep learning have achieved remarkable results on remote sensing images. However, the performance of the detector degrades when the test images are distinct from the training images. Domain adaptive vehicle detection is a promising approach to bridging the domain gap. Existing methods usually adopt fully shared networks, but ignore the problem that features from different domains may be incompatible within a single model. In this letter, we present a novel domain adaptive vehicle detection method based on patch-wise domain-specific channel recalibration (PDSCR). The PDSCR module routes the feature to the corresponding network branch and extracts the channel dependence using separate parameters. In this way, our method can explicitly capture domain-specific information for each domain. Furthermore, we propose a dynamic weighted prototype alignment (DWPA) to avoid the negative effects of false pseudo-labels, especially in the early stage of training. Experimental results of adaptation from our synthetic dataset to three real vehicle detection datasets demonstrate the effectiveness of our method. Code and our synthetic data will be available at https://weix-liu.github.io/.

Optical and synthetic aperture radar (SAR) image keypoint detection is an important foundation for multimodal remote sensing image matching. The influence of nonlinear radiometric differences and geometric deformation between optical and SAR images leads to low repeatability of existing keypoint detection methods. To address the problem that existing keypoint detection methods cannot provide the required homonymous points for heterogenous image matching, we propose a keypoint detection method (SKD-Net) for optical and SAR images, and improve it in terms of both network structure and network optimization. First, we propose a multilevel attention Siamese network, which is composed of multiple convolutional modules and transformer modules with shared weights to extract common features at different levels for keypoint detection. We introduce a transformer module in the keypoint detection pipeline and fuse shallow and deep features to obtain more spatial and rich semantic information to facilitate heterogeneous image keypoint detection. Then, to ensure that the detected keypoints have more homonymous points and localization accuracy, we propose a position consistent loss. Unlike previous loss functions, our designed position-consistent loss function takes the differences between heterogeneous image score maps into account, and it autonomously selects the optimized correct point pairs to enable the network to perform correct learning. Finally, extensive experiments show that our detection method outperforms the current state-of-the-art keypoint detection methods in terms of repeatability, localization accuracy, and matching performance. Our source code is available at https://github.com/zhangschen/ SKD-Net.

Deep learning approaches require enough training samples to perform well, but it is a challenge to collect enough real training data and label them manually. In this letter, we propose a practical framework for automatically generating content-rich synthetic images with ground-truth annotations. By rendering 3-D CAD models, we generate two synthetic aircraft image data sets with wide distribution (Syn N and Syn U). For improving the quality of synthetic images, we propose a multiscale attention module which enhances the Cycle-Consistent Adversarial Network (CycleGAN) in spatial and channel dimensions. Then, we compare the synthetic images before and after translation qualitatively and quantitatively. Experiments on Northwestern Polytechnical University (NWPU) very high resolution (VHR)-10, University of Chinese Academy of Sciences, orientation robust object detection in aerial images (UCAS-AOD), and benchmark for object DetectIon in Optical Remote sensing images (DIOR) data sets demonstrate that synthetic data augmentation can improve the performance of aircraft detection in remote sensing images, especially when real data are insufficient. Synthetic data are available at:https://weix-liu.github.io/.

In this paper, we design a Content-based map retrieval (CBM-R) system for land cover/land use (LCLU). The main contributions of our work are listed below. First, our system could allow the user to select region of interest as reference scene with variable shape and size. Whereas in traditional CBIR/CBMR systems, the region of interest is usually with fixed size of which is equal to the size of the analyse window for extracting features. In addition, the user could acquire various retrieval results by specifying corresponding parameters. In the end, by using combination of the feature library, the user could acquire the retrieval result faster.