Single Remote Sensing Research Articles

Single Remote Sensing Image Super-Resolution via a Generative Adversarial Network With Stratified Dense Sampling and Chain Training

Single Remote Sensing Image Super-Resolution via a Generative Adversarial Network With Stratified Dense Sampling and Chain Training

S2EFT: Spectral-Spatial-Elevation Fusion Transformer for hyperspectral image and LiDAR classification

With the innovation of sensors, communication, computer and other technologies, remote sensing (RS) imaging methods show a diversification trend. At present, relying on a variety of earth observation platforms, more multi-source RS data can be obtained, which overcomes the inadequacy and incompleteness of the new earth observation information of single source RS image. In this paper, a classification method of hyperspectral images (HSI) and LiDAR data based on spectral-spatial-elevation fusion Transformer (S2EFT) framework is proposed. The Transformer framework is introduced into the task of multi-source RS image classification. Two simple but effective modules, spatial information recognition module and sliding group spectral embedding module, are added to the proposed framework, and the Patch form commonly used in traditional convolutional neural networks (CNNs) is used as the input. It effectively solves the shortcomings of Transformer’s insufficient attention to local information, reduces redundant spatial information, enhances the transmission of information, and fully integrates multidimensional information of pixels in same position. Experiments on three real data sets are compared with existing methods, and the best results are obtained by the proposed methods. The source code can be downloaded from https://github.com/SYFYN0317/S2EFT.

Multi-Modal Remote Sensing Image Segmentation with Intuition-Inspired Hypergraph Modeling.

Multi-modal remote sensing (RS) image segmentation aims to comprehensively utilize multiple RS modalities to assign pixel-level semantics to the studied scenes, which can provide a new perspective for global city understanding. Multi-modal segmentation inevitably encounters the challenge of modeling intra- and inter-modal relationships, i.e., object diversity and modal gaps. However, the previous methods are usually designed for a single RS modality, limited by the noisy collection environment and poor discrimination information. Neuropsychology and neuroanatomy confirm that the human brain performs the guiding perception and integrative cognition of multi-modal semantics through intuitive reasoning. Therefore, establishing a semantic understanding framework inspired by intuition to realize multi-modal RS segmentation becomes the main motivation of this work. Drived by the superiority of hypergraphs in modeling high-order relationships, we propose an intuition-inspired hypergraph network (I2HN) for multi-modal RS segmentation. Specifically, we present a hypergraph parser to imitate guiding perception to learn intra-modal object-wise relationships. It parses the input modality into irregular hypergraphs to mine semantic clues and generate robust mono-modal representations. In addition, we also design a hypergraph matcher to dynamically update the hypergraph structure from the explicit correspondence of visual concepts, similar to integrative cognition, to improve cross-modal compatibility when fusing multi-modal features. Extensive experiments on two multi-modal RS datasets show that the proposed I2HN outperforms the state-of-the-art models, achieving F1/mIoU accuracy 91.4%/82.9% on the ISPRS Vaihingen dataset, and 92.1%/84.2% on the MSAW dataset. The complete algorithm and benchmark results will be available online.

Haze Removal for a Single Remote Sensing Image Using Low-Rank and Sparse Prior

Due to the influence of atmospheric scattering, the quality of remote sensing images is degraded, which severely limits the utility of remote sensing images. In this article, a novel dehazing algorithm for a single remote sensing image is proposed based on a low-rank and sparse prior (LSP). According to an atmospheric scattering model, the dark channel of a hazy image is decomposed into two parts: the dark channel of direct attenuation with sparseness and the atmospheric veil with low rank. The prior is obtained from the overall decomposition of the image rather than the patches of the image; therefore, the image pixel changes of the local blocks have little influence on the prior. Considering different resolutions of remote sensing images, the calculations of blocks involved in this article are completed by adaptive methods. The principal component pursuit and alternating direction multiplier method (PCP-ADMM) combined with the adaptive threshold shrinkage method are used for low-rank and sparse decomposition, therefore, the coarse estimation of the atmospheric veil is obtained. The guided filter with adaptive radius is used to refine it, and then the accurate atmospheric light is estimated. Finally, using the deformed atmospheric scattering model based on the atmospheric veil and atmospheric light, the haze-free image is restored. Extensive experimental results on publicly available data sets show that the dehazed images have abundant detail, high contrast, and minimal color distortion when using the proposed method, which is competitive with most state-of-the-art technologies.

Single Remote Sensing Image Dehazing Using Gaussian and Physics-Guided Process

Remote Sensing (RS) dehazing is a challenging task due to various haze distribution severely degrade the image quality. Recent learning-based methods achieve dramatic performance for RS dehazing, however previous ways are limited to their generality using only fully labeled datasets and less prior-guided information. In this paper, we explore the Gaussian and Physics-guided Dehazing Network (GPD-Net) to better obtain hazy feature and improve the generalization ability in real world condition. To promote the feature extraction, a novel Global Attention Mechanism (GAM) is involved to extract feature from different haze distribution. Then, an encoder-decoder network is designed with Gaussian Process (GP) in the intermediate latent space, in order to learn the full labeled dehazing and guide to handle the unlabeled learning. For the fine-tuning, we select some physical prior knowledge to refine the dehazed results. Extensive experiments demonstrate that our method outperforms the recent comparing approaches on the synthetic and real-world datasets.

Hybrid High-Resolution Learning for Single Remote Sensing Satellite Image Dehazing

Recently, deep learning models have shown convincing performance in removing a single satellite image haze, which arouses increasing attention in the field of remote sensing (RS). Unfortunately, these models still suffer from an insufficient ability to recover the desired fine spatial details from the hazy image. In this letter, we first attempt to explore an end-to-end hybrid high-resolution learning network framework termed H <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> RL-Net to address this issue due to its novel feature extraction architecture, where spatially precise outputs are guaranteed by the main high-resolution branch and semantically richer features are collected by the complementary set of multiresolution convolution streams. To improve representation learning, H <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> RL-Net is constructed primarily by exploiting the parallel cross-scale fusion (PCF) module, thereby increasingly aggregating information from the multiple scales at the respective resolution level, which allows both top-down and bottom-up information exchanging processes. Simultaneously, we also introduce the channel feature refinement (CFR) block to our model, aiming to perform dynamic feature recalibration among the channelwise features and produce better dehazed results. The experimental analysis illustrates that the designed framework can deliver significant improvements over other baseline methods in the synthetic and real-world hazy RS images under various scenes.

Statistical evaluation of on-road vehicle emissions measurement using a dual remote sensing technique

On-road remote sensing (RS) is a rapid, non-intrusive and economical tool to monitor and control the emissions of in-use vehicles, and currently is gaining popularity globally. However, a majority of studies used a single RS technique, which may bias the measurements since RS only captures a snapshot of vehicle emissions. This study aimed to use a unique dual RS technique to assess the characteristics of on-road vehicle emissions. The results show that instantaneous vehicle emissions are highly dynamic under real-world driving conditions. The two emission factors measured by the dual RS technique show little correlation, even under the same driving condition. This indicates that using the single RS technique may be insufficient to accurately represent the emission level of a vehicle based on one measurement. To increase the accuracy of identifying high-emitting vehicles, using the dual RS technique is essential. Despite little correlation, the dual RS technique measures the same average emission factors as the single RS technique does when a large number of measurements are available. Statistical analysis shows that both RS systems demonstrate the same Gamma distribution with ≥200 measurements, leading to converged mean emission factors for a given vehicle group. These findings point to the need for a minimum sample size of 200 RS measurements in order to generate reliable emission factors for on-road vehicles. In summary, this study suggests that using the single or dual RS technique will depend on the purpose of applications. Both techniques have the same accuracy in calculating average emission factors when sufficient measurements are available, while the dual RS technique is more accurate in identifying high-emitters based on one measurement only.

Blind Super-Resolution for Single Remote Sensing Image via Sparse Representation and Transformed Self-Similarity

Super-resolution (SR) reconstruction is one of the effective ways to improve the spatial resolution of remote sensing images, but the blur kernel estimation without any additional prior information is the key to the quality of SR reconstruction. Facing the above problem, a blind super-resolution method via sparse representation and transformed self-similarity is proposed in this paper. In this method, the blur kernel as well as the reconstructed high-resolution image are estimated at the same time using the structural self-similarity from a single image, a down-sampled version of the observed image is used as the training samples for the dictionary learning, and the internal patch search space is expanded using the transformed self-similarity to improve the quality of estimation. Experiment results show that, our method performs well both in quality of kernel estimation and SR reconstruction.

Spatial Perspectives toward the Recommendation of Remote Sensing Images Using the INDEX Indicator, Based on Principal Component Analysis

Progress in the development of sensor technology has increased the speed and convenience of remote sensing (RS) image acquisition. As the volume of RS images steadily increases, the challenge is no longer in producing and acquiring an RS image, but in finding a particular image from numerous RS images that precisely meets user application needs. Some spatial measuring methods specific to the recommendation of RS images have been proposed and could be used to score and sort RS images according to users’ requests. Our previous study introduced two measuring methods, namely, available space (AS) and image extension (IE), which have similar results but complementary effects for spatially ranking recommended images. The AS indicator could cover the inadequacies of the IE indicator in some cases and vice versa. The current study combines these two indicators using principal component analysis and produced a new indicator called INDEX, which we used in the RS image spatial recommendation. The ranking results were measured using a normalized discounted cumulative gain (NDCG) and several other statistic criteria. The results indicate that users are more satisfied with the recommendations of the INDEX indicator than those of AS, IE and Hausdorff distance for single RS image type selections which is the most common scenario for RS image applications. When dealing with hybrid RS image types, the INDEX indicator performs very closely to the dominant IE indicator, yet maintaining the characteristics of the AS indicator.

Single Remote Sensing Image Super-Resolution with an Adaptive Joint Constraint Model.

Remote sensing images have been widely used in many applications. However, the resolution of the obtained remote sensing images may not meet the increasing demands for some applications. In general, the sparse representation-based super-resolution (SR) method is one of the most popular methods to solve this issue. However, traditional sparse representation SR methods do not fully exploit the complementary constraints of images. Therefore, they cannot accurately reconstruct the unknown HR images. To address this issue, we propose a novel adaptive joint constraint (AJC) based on sparse representation for the single remote sensing image SR. First, we construct a nonlocal constraint by using the nonlocal self-similarity. Second, we propose a local structure filter according to the local gradient of the image and then construct a local constraint. Next, the nonlocal and local constraints are introduced into the sparse representation-based SR framework. Finally, the parameters of the joint constraint model are selected adaptively according to the level of image noise. We utilize the alternate iteration algorithm to tackle the minimization problem in AJC. Experimental results show that the proposed method achieves good SR performance in preserving image details and significantly improves the objective evaluation indices.

Haze removal from a single remote sensing image based on a fully convolutional neural network

In many remote sensing (RS) applications, haze greatly affects the quality of optical RS images, but we do not always have the conditions to acquire multiple images in the same area for haze removal tasks. Therefore, the research on haze removal from a single RS image is necessary. Previous haze-removal methods introduce various prior knowledge to solve this problem, and thus, the quality of these methods largely depends on the reliability and validity of prior knowledge, which brings various limitations. We propose and validate a deep-learning-based model for haze removal, named haze removal fully convolutional network, to estimate transmission maps and generate corresponding haze-removed images via an atmospheric scattering model. Moreover, we propose an approximate method to produce hazy-and-clear image pairs as a dataset for training and validation. Experiments using this dataset demonstrated that the proposed model achieved the desired results in both visual effect and quantitative measurement.

Single Remote Sensing Multispectral Image Dehazing Based on a Learning Framework

Given that a single remote sensing image dehazing is an ill-posed problem, this is still a challenging task. In order to improve the visibility of a single hazy remote sensing multispectral image, we developed a novel and effective algorithm based on a learning framework. A linear regression model with the relevant features of haze was established. And the gradient descent method is applied to the learning model. Then a hazy image accurate transmission map is obtained by learning the coefficients of the linear model. In addition, we proposed a more effective method to estimate the atmospheric light, which can restrain the influence of highlight areas on the atmospheric light acquisition. Compared with the traditional haze removal methods, the experimental results demonstrate that the proposed algorithm can achieve better visual effect and color fidelity. Both subjective evaluation and objective assessments indicate that the proposed method achieves a better performance than the state-of-the-art methods.

Haze Removal for a Single Remote Sensing Image Based on Deformed Haze Imaging Model

The contrast of remote sensing images captured in haze condition is poor, which influences their interpretation. In this letter, a novel dehazing algorithm based on the deformed haze imaging model is proposed. First, the model is deformed by introducing a translation term. Second, the atmospheric light and transmission are estimated according to the new model combined with dark channel prior. Lastly, the haze is successfully removed from remote sensing images using the proposed estimation algorithm. The estimated transmission is insensitive to the texture of ground objects, and the dehazing effect for nonuniform haze is more satisfactory than the compared method. Moreover, our approach can be used for general haze removal through adjusting the translation term. Experimental results reveal that the proposed method can recover the real scene clearly from haze remote sensing images along with the advantage of good color consistency.

A Stochastic Framework for the Identification of Building Rooftops Using a Single Remote Sensing Image

The identification of building rooftops from a single image, without the use of auxiliary 3-D information like stereo pairs or digital elevation models, is a very challenging and difficult task in the area of remote sensing. The existing methodologies rarely tackle the problem of 3-D object identification, like buildings, from a purely stochastic viewpoint. Our approach is based on a stochastic image interpretation model, which combines both 2-D and 3-D contextual information of the imaged scene. Building rooftop hypotheses are extracted using a contour-based grouping hierarchy that emanates from the principles of perceptual organization. We use a Markov random field model to describe the dependencies between all available hypotheses with regard to a globally consistent interpretation. The hypothesis verification step is treated as a stochastic optimization process that operates on the whole grouping hierarchy to find the globally optimal configuration for the locally interacting grouping hypotheses, providing also an estimate of the height of each extracted rooftop. This paper describes the main principles of our method and presents building detection results on a set of synthetic and airborne images.