Haze superimposes a veil over remote sensing images, which severely limits the extraction of valuable military information. To this end, we present a novel trinity model to restore realistic surface information by integrating the merits of both prior-based and deep learning-based strategies. Concretely, the critical insight of our Trinity-Net is to investigate how to incorporate prior information into CNNs and Swin Transformer for reasonable estimation of haze parameters. Then, haze-free images are obtained by reconstructing the remote sensing image formation model. Although Swin Transformer has shown tremendous potential in the dehazing task, which typically results in ambiguous details. We devise a gradient guidance module that naturally inherits structure priors of gradient maps, guiding the deep model to generate visually pleasing details. In light of the generality of image formation parameters, we successfully promote Trinity-Net to natural image dehazing and underwater image enhancement tasks. Notably, the acquisition of large-scale remote sensing hazy images and natural hazy images in military scenes is not feasible in practice. To bridge this gap, we construct a Remote Sensing Image Dehazing Benchmark (RSID) and a Natural Image Dehazing Benchmark (NID), including 1000 real-world hazy images with corresponding ground truth images, respectively. To our knowledge, this is the first exploration to develop dehazing benchmarks in the military field, alleviating the dilemma of data scarcity. Extensive experiments on three vision tasks illustrate the superiority of our Trinity-Net against multiple state-of-the-art methods. The datasets and code are available at https://github.com/chi-kaichen/Trinity-Net.

The objective of semantic change detection (SCD) is to discern intricate changes in land cover while simultaneously identifying their semantic categories. Prior research has shown that using multiple independent branches for the distinct tasks of change localization and semantic recognition is a reliable approach to solving the SCD problem. Nevertheless, conventional SCD architectures rely heavily on a high degree of consistency within the bitemporal feature space when modeling difference features, inevitably resulting in false positives or missed alerts within change areas. In this article, we introduce an SCD framework called the cross-differential semantic consistency network. Cross-difference semantic consistency (CdSC) is designed to mine deep discrepancies in bitemporal instance features while preserving their semantic consistency. Specifically, the 3-D cross-difference module, incorporating 3-D convolutions, explores the interaction of cross-temporal features, revealing inherent differences among various land features. Simultaneously, deep semantic representations are further utilized to enhance the local correlation of difference information, thereby improving the model’s discriminative capabilities within change regions. Incorporating principles from contrastive learning, a semantic co-alignment (SCA) loss is introduced to increase intra-class consistency and inter-class distinctiveness of dual-temporal semantic features, thereby addressing the challenges posed by semantic disparities. Extensive experiments on two SCD datasets demonstrate that CdSC outperforms other state-of-the-art SCD methods significantly in both qualitative and quantitative evaluations. The code and dataset are available at https://github.com/weiAI1996/CdSC .

Since light is scattered and absorbed by water, underwater images have inherent degradation (e.g. hazing, color shift), consequently impeding the development of remotely operated vehicles. Towards this end, we propose a novel method, referred to as Best o f Bo th Worlds (Boths). With parameters of only 0.0064M, Boths can be considered a super lightweight neural network for underwater image enhancement. On the whole, it has three levels: Structure and Detail Features; Pixel and Channel Dimensions; High and Low Frequency Information. Each of these three levels represents ”Best of Both Worlds”. Initially, by interacting with structure and detail features, Boths can focus on these two aspects at the same time. Further, our network can simultaneously consider channel and pixel dimensions through 3D attention learning, which is more similar to human visual perception. Lastly, the proposed model can focus on high and low frequency information, through a novel loss function based on the wavelet transforms. Upon subsequent analysis and evaluation, Boths has shown superior performance compared with state-of-the-art methods. Our models and datasets are publicly available at: https://github.com/perseveranceLX/Boths.

Shadow removal is an essential task for remote sensing imagery analysis, which is tricky due to spatial irregular and inhomogeneous degradation distribution. Unfortunately, current shadow removal pipelines face challenges with suboptimal performance and insufficient interpretability. To this end, we unleash the long-sequence modeling potential of State Space Models (SSMs) in the context of shadow removal. Coupled with the accurate perception of traditional Retinex decomposition towards illumination, the well-designed RSMamba enjoys the best of both worlds between superior competitiveness and theoretical intuitiveness. Specifically, RSMamba mimics the retina and cerebral cortex to explore illumination and reflectance. The former drives the selective scan mechanism to enhance the response towards contamination, while the latter serves as a tool to preserve illumination fidelity. In addition, contour and gradient regularizations of illumination and reflectance components reflect the spatial opponency of shadows, which are consistent with the center-surround opponent receptive field of the human visual system. Such a manner incorporates the domain knowledge of neurophysiological mechanisms into neural networks, providing new insights into shadow removal. Extensive experiments demonstrate that RSMamba outperforms state-of-the-art methods.

Remote sensing shadow removal is an open issue. Previous studies focus on working in the spatial dimension, ignoring the potential of the Fourier dimension, while illumination degradation typically exists in the amplitude component. To address this limitation, our insight is a fresh dual-stage Fourier-based network (NeFour), which explores the best of both worlds between frequency and spatial information. In the frequency stage, we investigate the positive correlation between amplitude and brightness from channel and spatial statistics. Coupled with implicitly defined normalization, a controllable fitting amplitude transform map recreates the illumination. In the spatial stage, the inverted dark channel prior with 3D coordinates serves as modulation matrices that naturally reveal the spatial distribution of shadows, thus elegantly eliminating shadow remnants. With ingenious design, NeFour achieves nontrivial performance against state-of-the-art shadow removal methods in terms of both visual perception and quantitative evaluation. The code is publicly available at https://github.com/chi-kaichen/NeFour.

Light can be absorbed and scattered when traveling through water, which results in underwater optical images suffering from blurring and color distortion. To improve the visual quality of underwater optical images, we propose a novel, to the best of our knowledge, image sharpening method. We utilize the relative total variation model to decompose images into structure and texture layers in a novel manner. On those two layers, the red-blue dark channel prior (RBDCP) and detail lifting algorithms are proposed, respectively. The RBDCP model calculates background light based on brightness, gradient discrimination, and hue judgment, which then generates transmission maps using red-blue channel attenuation characteristics. The linear combination of the Gaussian kernel and binary mask is employed in the proposed detail lifting algorithm. Furthermore, we combine the layers of restoration structure and enhancement texture for image sharpening, inspired by the concept of fusion. Our methodology has rich texture information and is effective in color correction and atomization removal through RBDCP. Extensive experimental results indicate that the proposed method effectively balances image hue, saturation, and clarity.