Remote Sensing Image Scene Research Articles

Searching for CNN Architectures for Remote Sensing Scene Classification

Convolutional neural network (CNN) models for remote sensing (RS) scene classification are largely built on pretrained networks that are trained on the general-purpose ImageNet dataset in computer vision. The pretrained networks can easily be adapted for transfer learning in RS scene classification. However, the accuracy of transfer learning may decline as RS images are considerably different from other images. Thus, the pretrained CNN model learned on ImageNet may not be sufficient for the accurate classification of RS image scenes. Furthermore, most of the pretrained models have large memory footprints, which place a further burden on computational requirements. In this work, we explore SLGE-based random search with early stopping in the search for CNN architectures for both single-label and multilabel RS scene classification tasks. In SLGE, the architecture search space is capable of representing multipath Inception-like modular cells with skip-connections similar to human-expert designs. The experimental results on four RS scene classification benchmarks show that the automatically discovered networks demonstrate the promising capability in classifying multispectral satellite image scenes compared with fine-tuned pretrained CNN models. Using fewer parameters with 0.56B FLOPS, our best network achieves a classification accuracy rate of 96.56% and 96.10% on NWPU-RESISC45 single-label and AID single-label RGB aerial image datasets, respectively, and the classification accuracy rate of 99.76% and 93.89% on EuroSAT single-label and BigEarthNet multilabel multispectral satellite image datasets, respectively. The results position our approach among the best of the state of the art.

Multilayer Feature Fusion Network With Spatial Attention and Gated Mechanism for Remote Sensing Scene Classification

Remote sensing (RS) scene classification has attracted extensive attention due to its large number of applications. Recently, convolutional neural networks (CNNs) methods have shown impressive ability of feature learning in RS scene classification. However, the performance is still limited by large-scale variance and complex background. To address these problems, we present a multilayer feature fusion network with spatial attention and gated mechanism (MLF2Net_SAGM) for RS scene classification. At first, the backbone is employed to extract multilayer convolutional features. Then, a residual spatial attention module (RSAM) is proposed to enhance discriminative regions of the multilayer feature maps, and key areas can be harvested. Finally, the multilayer spatial calibration features are fused to form the final feature map, and a gated fusion module (GFM) is designed to eliminate feature redundancy and mutual exclusion (FRME). To verify the effectiveness of the proposed method, we conduct comparative experiments based on three widely used RS image scene classification benchmarks. The results show that the direct fusion of multilayer features via element-wise addition leads to FRME, whereas our method fuses multilayer features more effectively and improves the performance of scene classification.

RDPN: Tackling the Data Shift Problem in Remote Sensing Scene Classification

Currently, remote-sensing (RS) scene classification has played an important role in many practical applications. However, traditional methods that are based on deep convolutional neural networks (DCNNs) have several difficulties when faced with data shift problems: novel classes, varied orientations, and large intraclass variations of RS scene images. In this letter, we propose the rotation-invariant and discriminative-learning prototypical networks (RDPNs) for RS scene classification. RDPN uses A-ORConv32 basic blocks and attention mechanisms to obtain rotation-invariant and discriminative features. In addition, adaptive cosine center loss is proposed to constrain the features to mitigate the large intraclass variations and penalize the hard samples adaptively. We conduct extensive experiments on publicly available datasets and achieve 1.69%–19.38% higher accuracy than existing methods. The experimental results verify that the proposed RDPN can solve the data shift problems well in RS scene classification.

A Lightweight Convolutional Neural Network Based on Group-Wise Hybrid Attention for Remote Sensing Scene Classification

With the development of computer vision, attention mechanisms have been widely studied. Although the introduction of an attention module into a network model can help to improve classification performance on remote sensing scene images, the direct introduction of an attention module can increase the number of model parameters and amount of calculation, resulting in slower model operations. To solve this problem, we carried out the following work. First, a channel attention module and spatial attention module were constructed. The input features were enhanced through channel attention and spatial attention separately, and the features recalibrated by the attention modules were fused to obtain the features with hybrid attention. Then, to reduce the increase in parameters caused by the attention module, a group-wise hybrid attention module was constructed. The group-wise hybrid attention module divided the input features into four groups along the channel dimension, then used the hybrid attention mechanism to enhance the features in the channel and spatial dimensions for each group, then fused the features of the four groups along the channel dimension. Through the use of the group-wise hybrid attention module, the number of parameters and computational burden of the network were greatly reduced, and the running time of the network was shortened. Finally, a lightweight convolutional neural network was constructed based on the group-wise hybrid attention (LCNN-GWHA) for remote sensing scene image classification. Experiments on four open and challenging remote sensing scene datasets demonstrated that the proposed method has great advantages, in terms of classification accuracy, even with a very low number of parameters.

LPIN: A Lightweight Progressive Inpainting Network for Improving the Robustness of Remote Sensing Images Scene Classification

At present, the classification accuracy of high-resolution Remote Sensing Image Scene Classification (RSISC) has reached a quite high level on standard datasets. However, when coming to practical application, the intrinsic noise of satellite sensors and the disturbance of atmospheric environment often degrade real Remote Sensing (RS) images. It introduces defects to them, which affects the performance and reduces the robustness of RSISC methods. Moreover, due to the restriction of memory and power consumption, the methods also need a small number of parameters and fast computing speed to be implemented on small portable systems such as unmanned aerial vehicles. In this paper, a Lightweight Progressive Inpainting Network (LPIN) and a novel combined approach of LPIN and the existing RSISC methods are proposed to improve the robustness of RSISC tasks and satisfy the requirement of methods on portable systems. The defects in real RS images are inpainted by LPIN to provide a purified input for classification. With the combined approach, the classification accuracy on RS images with defects can be improved to the original level of those without defects. The LPIN is designed on the consideration of lightweight model. Measures are adopted to ensure a high gradient transmission efficiency while reducing the number of network parameters. Multiple loss functions are used to get reasonable and realistic inpainting results. Extensive tests of image inpainting of LPIN and classification tests with the combined approach on NWPU-RESISC45, UC Merced Land-Use and AID datasets are carried out which indicate that the LPIN achieves a state-of-the-art inpainting quality with less parameters and a faster inpainting speed. Furthermore, the combined approach keeps the comparable classification accuracy level on RS images with defects as that without defects, which will improve the robustness of high-resolution RSISC tasks.

A Lightweight Convolutional Neural Network Based on Channel Multi-Group Fusion for Remote Sensing Scene Classification

With the development of remote sensing scene image classification, convolutional neural networks have become the most commonly used method in this field with their powerful feature extraction ability. In order to improve the classification performance of convolutional neural networks, many studies extract deeper features by increasing the depth and width of convolutional neural networks, which improves classification performance but also increases the complexity of the model. To solve this problem, a lightweight convolutional neural network based on channel multi-group fusion (LCNN-CMGF) is presented. For the proposed LCNN-CMGF method, a three-branch downsampling structure was designed to extract shallow features from remote sensing images. In the deep layer of the network, the channel multi-group fusion structure is used to extract the abstract semantic features of remote sensing scene images. The structure solves the problem of lack of information exchange between groups caused by group convolution through channel fusion of adjacent features. The four most commonly used remote sensing scene datasets, UCM21, RSSCN7, AID and NWPU45, were used to carry out a variety of experiments in this paper. The experimental results under the conditions of four datasets and multiple training ratios show that the proposed LCNN-CMGF method has more significant performance advantages than the compared advanced method.

A Semantic-Preserving Deep Hashing Model for Multi-Label Remote Sensing Image Retrieval

Conventional remote sensing image retrieval (RSIR) systems perform single-label retrieval with a single label to represent the most dominant semantic content for an image. Improved spatial resolution dramatically boosts the remote sensing image scene complexity, as a remote sensing image always contains multiple categories of surface features. In this case, a single label cannot comprehensively describe the semantic content of a complex remote sensing image scene and therefore results in poor retrieval performance in practical applications. As a result, researchers have begun to pay attention to multi-label image retrieval. However, in the era of massive remote sensing data, how to increase retrieval efficiency and reduce feature storage while preserving semantic information remains unsolved. Considering the powerful capability of hashing learning in overcoming the curse of dimensionality caused by high-dimensional image representation in Approximate Nearest Neighbor (ANN) search problems, we propose a new semantic-preserving deep hashing model for multi-label remote sensing image retrieval. Our model consists of three main components: (1) a convolutional neural network to extract image features; (2) a hash layer to generate binary codes; (3) a new loss function to better maintain the multi-label semantic information of hash learning contained in context remote sensing image scene. As far as we know, this is the first attempt to apply deep hashing into the multi-label remote sensing image retrieval. Experimental results indicate the effectiveness and promising of the introduction of hashing methods in the multi-label remote sensing image retrieval.

Review of Image Classification Algorithms Based on Convolutional Neural Networks

Image classification has always been a hot research direction in the world, and the emergence of deep learning has promoted the development of this field. Convolutional neural networks (CNNs) have gradually become the mainstream algorithm for image classification since 2012, and the CNN architecture applied to other visual recognition tasks (such as object detection, object localization, and semantic segmentation) is generally derived from the network architecture in image classification. In the wake of these successes, CNN-based methods have emerged in remote sensing image scene classification and achieved advanced classification accuracy. In this review, which focuses on the application of CNNs to image classification tasks, we cover their development, from their predecessors up to recent state-of-the-art (SOAT) network architectures. Along the way, we analyze (1) the basic structure of artificial neural networks (ANNs) and the basic network layers of CNNs, (2) the classic predecessor network models, (3) the recent SOAT network algorithms, (4) comprehensive comparison of various image classification methods mentioned in this article. Finally, we have also summarized the main analysis and discussion in this article, as well as introduce some of the current trends.

Remote Sensing Image Scene Classification Based on Global Self-Attention Module

The complexity of scene images makes the research on remote-sensing image scene classification challenging. With the wide application of deep learning in recent years, many remote-sensing scene classification methods using a convolutional neural network (CNN) have emerged. Current CNN usually output global information by integrating the depth features extricated from the convolutional layer through the fully connected layer; however, the global information extracted is not comprehensive. This paper proposes an improved remote-sensing image scene classification method based on a global self-attention module to address this problem. The global information is derived from the depth characteristics extracted by the CNN. In order to better express the semantic information of the remote-sensing image, the multi-head self-attention module is introduced for global information augmentation. Meanwhile, the local perception unit is utilized to improve the self-attention module’s representation capabilities for local objects. The proposed method’s effectiveness is validated through comparative experiments with various training ratios and different scales on public datasets (UC Merced, AID, and NWPU-NESISC45). The precision of our proposed model is significantly improved compared to other methods for remote-sensing image scene classification.

Remote Sensing Scene Image Classification Based on Dense Fusion of Multi-level Features

For remote sensing scene image classification, many convolution neural networks improve the classification accuracy at the cost of the time and space complexity of the models. This leads to a slow running speed for the model and cannot realize a trade-off between the model accuracy and the model running speed. As the network deepens, it is difficult to extract the key features with a sample double branched structure, and it also leads to the loss of shallow features, which is unfavorable to the classification of remote sensing scene images. To solve this problem, we propose a dual branch multi-level feature dense fusion-based lightweight convolutional neural network (BMDF-LCNN). The network structure can fully extract the information of the current layer through 3 × 3 depthwise separable convolution and 1 × 1 standard convolution, identity branches, and fuse with the features extracted from the previous layer 1 × 1 standard convolution, thus avoiding the loss of shallow information due to network deepening. In addition, we propose a downsampling structure that is more suitable for extracting the shallow features of the network by using the pooled branch to downsample and the convolution branch to compensate for the pooled features. Experiments were carried out on four open and challenging remote sensing image scene data sets. The experimental results show that the proposed method has higher classification accuracy and lower model complexity than some state-of-the-art classification methods and realizes the trade-off between model accuracy and model running speed.

DRSNet: Novel architecture for small patch and low-resolution remote sensing image scene classification

Mainstream architectures of convolutional neural networks (CNNs) for image recognition follow the protocol of shrinking input data size to extract semantic information. However, poor overall accuracy (OA) may be achieved when dealing with small input size and low-resolution images with classic CNNs. This paper proposes a novel, deep CNN architecture called DRSNet for small patch size Landsat 8 remote sensing (RS) image recognition. A module called residual inception channel attention block is applied for feature extraction, which combines the advantages of Inception-ResNet and channel attention. Pooling layers are replaced with reduction modules to prevent representational bottleneck. Unlike existing CNN structures, DRSNet adopts upsampling steps before final pooling layers, retrieving lost information caused by previous downsampling steps. Compared with existing state-of-the-art CNNs, DRSNet achieves the highest classification accuracy in our RS data set. Experimental results show that our model improves OA by 2%–9%, accelerates convergence speed, and effectively reduces loss. DRSNet also exhibits impressive results and outperforms baseline CNNs in three public data sets, namely, EuroSAT, Brazilian Coffee Scenes, and UCMerced Land Use. The proposed network is practical for large-scale land surface classification using free, low-resolution RS data; hence, it could be useful for nongovernment organizations or governments of developing countries.

Early Labeled and Small Loss Selection Semi-Supervised Learning Method for Remote Sensing Image Scene Classification

The classification of aerial scenes has been extensively studied as the basic work of remote sensing image processing and interpretation. However, the performance of remote sensing image scene classification based on deep neural networks is limited by the number of labeled samples. In order to alleviate the demand for massive labeled samples, various methods have been proposed to apply semi-supervised learning to train the classifier using labeled and unlabeled samples. However, considering the complex contextual relationship and huge spatial differences, the existing semi-supervised learning methods bring different degrees of incorrectly labeled samples when pseudo-labeling unlabeled data. In particular, when the number of labeled samples is small, it affects the generalization performance of the model. In this article, we propose a novel semi-supervised learning method with early labeled and small loss selection. First, the model learns the characteristics of simple samples in the early stage and uses multiple early models to screen out a small number of unlabeled samples for pseudo-labeling based on this characteristic. Then, the model is trained in a semi-supervised manner by combining labeled samples, pseudo-labeled samples, and unlabeled samples. In the training process of the model, small loss selection is used to further eliminate some of the noisy labeled samples to improve the recognition accuracy of the model. Finally, in order to verify the effectiveness of the proposed method, it is compared with several state-of-the-art semi-supervised classification methods. The results show that when there are only a few labeled samples in remote sensing image scene classification, our method is always better than previous methods.

Semantic Multigranularity Feature Learning for High-Resolution Remote Sensing Image Scene Classification

High-resolution remote sensing image scene classification is a challenging visual task due to the large intravariance and small intervariance between the categories. To accurately recognize the scene categories, it is essential to learn discriminative features from both global and local critical regions. Recent efforts focus on how to encourage the network to learn multigranularity features with the destruction of the spatial information on the input image at different scales, which leads to meaningless edges that are harmful to training. In this study, we propose a novel method named Semantic Multigranularity Feature Learning Network (SMGFL-Net) for remote sensing image scene classification. The core idea is to learn both global and multigranularity local features from rearranged intermediate feature maps, thus, eliminating the meaningless edges. These features are then fused for the final prediction. Our proposed framework is compared with a collection of state-of-the-art (SOTA) methods on two fine-grained remote sensing image scene datasets, including the NWPU-RESISC45 and Aerial Image Datasets (AID). We justify several design choices, including the branch granularities, fusion strategies, pooling operations, and necessity of feature map rearrangement through a comparative study. Moreover, the overall performance results show that SMGFL-Net consistently outperforms other peer methods in classification accuracy, and the superiority is more apparent with less training data, demonstrating the efficacy of feature learning of our approach.

Decision-Level Fusion with a Pluginable Importance Factor Generator for Remote Sensing Image Scene Classification

Remote sensing image scene classification acts as an important task in remote sensing image applications, which benefits from the pleasing performance brought by deep convolution neural networks (CNNs). When applying deep models in this task, the challenges are, on one hand, that the targets with highly different scales may exist in the image simultaneously and the small targets could be lost in the deep feature maps of CNNs; and on the other hand, the remote sensing image data exhibits the properties of high inter-class similarity and high intra-class variance. Both factors could limit the performance of the deep models, which motivates us to develop an adaptive decision-level information fusion framework that can incorporate with any CNN backbones. Specifically, given a CNN backbone that predicts multiple classification scores based on the feature maps of different layers, we develop a pluginable importance factor generator that aims at predicting a factor for each score. The factors measure how confident the scores in different layers are with respect to the final output. Formally, the final score is obtained by a class-wise and weighted summation based on the scores and the corresponding factors. To reduce the co-adaptation effect among the scores of different layers, we propose a stochastic decision-level fusion training strategy that enables each classification score to randomly participate in the decision-level fusion. Experiments on four popular datasets including the UC Merced Land-Use dataset, the RSSCN 7 dataset, the AID dataset, and the NWPU-RESISC 45 dataset demonstrate the superiority of the proposed method over other state-of-the-art methods.

An Empirical Study of Adversarial Examples on Remote Sensing Image Scene Classification

Deep neural networks (DNNs), which learn a hierarchical representation of features, have shown remarkable performance in big data analytics of remote sensing. However, previous research indicates that DNNs are easily spoofed by adversarial examples, which are crafted images with artificial perturbations that fool DNN models toward wrong predictions. To comprehensively evaluate the impact of adversarial examples on the remote sensing image (RSI) scene classification, this study tests eight state-of-the-art classification DNNs on six RSI benchmarks. These data sets include both optical and synthetic-aperture radar (SAR) images of different spectral and spatial resolutions. In the experiment, we create 48 classification scenarios and use four cutting-edge attack algorithms to investigate the influence of the adversarial example on the classification of RSIs. The experimental result shows that the fooling rates of the attacks are all over 98% across the 48 scenarios. We also find that, for the optical data, the seriousness of the adversarial problem has a negative relationship with the richness of the feature information. Besides, adversarial examples generated from SAR images are used easily for fooling the models with an average fooling rate of 76.01%. By analyzing the class distribution of these adversarial examples, we find that the distribution of the misclassifications is not affected by the types of models and attack algorithms-adversarial examples of RSIs of the same class cluster on fixed several classes. The analysis of classes of adversarial examples not only helps us explore the relationships between data set classes but also provides insights for further designing defensive algorithms.

Robust deep alignment network with remote sensing knowledge graph for zero-shot and generalized zero-shot remote sensing image scene classification

Although deep learning has revolutionized remote sensing (RS) image scene classification, current deep learning-based approaches highly depend on the massive supervision of predetermined scene categories and have disappointingly poor performance on new categories that go beyond predetermined scene categories. In reality, the classification task often has to be extended along with the emergence of new applications that inevitably involve new categories of RS image scenes, so how to make the deep learning model own the inference ability to recognize the RS image scenes from unseen categories, which do not overlap the predetermined scene categories in the training stage, becomes incredibly important. By fully exploiting the RS domain characteristics, this paper constructs a new remote sensing knowledge graph (RSKG) from scratch to support the inference recognition of unseen RS image scenes. To improve the semantic representation ability of RS-oriented scene categories, this paper proposes to generate a Semantic Representation of scene categories by representation learning of RSKG (SR-RSKG). To pursue robust cross-modal matching between visual features and semantic representations, this paper proposes a novel deep alignment network (DAN) with a series of well-designed optimization constraints, which can simultaneously address zero-shot and generalized zero-shot RS image scene classification. Extensive experiments on one merged RS image scene dataset, which is the integration of multiple publicly open datasets, show that the recommended SR-RSKG obviously outperforms the traditional knowledge types (e.g., natural language processing models and manually annotated attribute vectors), and our proposed DAN shows better performance compared with the state-of-the-art methods under both the zero-shot and generalized zero-shot RS image scene classification settings. The constructed RSKG will be made publicly available along with this paper (https://github.com/kdy2021/SR-RSKG).

An Attention-Guided Multilayer Feature Aggregation Network for Remote Sensing Image Scene Classification

Remote sensing image scene classification (RSISC) has broad application prospects, but related challenges still exist and urgently need to be addressed. One of the most important challenges is how to learn a strong discriminative scene representation. Recently, convolutional neural networks (CNNs) have shown great potential in RSISC due to their powerful feature learning ability; however, their performance may be restricted by the complexity of remote sensing images, such as spatial layout, varying scales, complex backgrounds, category diversity, etc. In this paper, we propose an attention-guided multilayer feature aggregation network (AGMFA-Net) that attempts to improve the scene classification performance by effectively aggregating features from different layers. Specifically, to reduce the discrepancies between different layers, we employed the channel–spatial attention on multiple high-level convolutional feature maps to capture more accurately semantic regions that correspond to the content of the given scene. Then, we utilized the learned semantic regions as guidance to aggregate the valuable information from multilayer convolutional features, so as to achieve stronger scene features for classification. Experimental results on three remote sensing scene datasets indicated that our approach achieved competitive classification performance in comparison to the baselines and other state-of-the-art methods.

Unsupervised Representation High-Resolution Remote Sensing Image Scene Classification via Contrastive Learning Convolutional Neural Network

Inspired by the outstanding achievement of deep learning, supervised deep learning representation methods for high-spatial-resolution remote sensing image scene classification obtained state-of-the-art performance. However, supervised deep learning representation methods need a considerable amount of labeled data to capture class-specific features, limiting the application of deep learning-based methods while there are a few labeled training samples. An unsupervised deep learning representation, high-resolution remote sensing image scene classification method is proposed in this work to address this issue. The proposed method, called contrastive learning, narrows the distance between positive views: color channels belonging to the same images widens the gaps between negative view pairs consisting of color channels from different images to obtain class-specific data representations of the input data without any supervised information. The classifier uses extracted features by the convolutional neural network (CNN)-based feature extractor with labeled information of training data to set space of each category and then, using linear regression, makes predictions in the testing procedure. Comparing with existing unsupervised deep learning representation high-resolution remote sensing image scene classification methods, contrastive learning CNN achieves state-of-the-art performance on three different scale benchmark data sets: small scale RSSCN7 data set, midscale aerial image data set, and large-scale NWPU-RESISC45 data set.

Prototype Calibration with Feature Generation for Few-Shot Remote Sensing Image Scene Classification

Few-shot classification of remote sensing images has attracted attention due to its important applications in various fields. The major challenge in few-shot remote sensing image scene classification is that limited labeled samples can be utilized for training. This may lead to the deviation of prototype feature expression, and thus the classification performance will be impacted. To solve these issues, a prototype calibration with a feature-generating model is proposed for few-shot remote sensing image scene classification. In the proposed framework, a feature encoder with self-attention is developed to reduce the influence of irrelevant information. Then, the feature-generating module is utilized to expand the support set of the testing set based on prototypes of the training set, and prototype calibration is proposed to optimize features of support images that can enhance the representativeness of each category features. Experiments on NWPU-RESISC45 and WHU-RS19 datasets demonstrate that the proposed method can yield superior classification accuracies for few-shot remote sensing image scene classification.

Classification of Remote Sensing Image Scenes Using Double Feature Extraction Hybrid Deep Learning Approach

Over the last decade, remote sensing technology has advanced dramatically, resulting in significant improvements on image quality, data volume, and application usage. These images have essential applications since they can help with quick and easy interpretation. Many standard detection algorithms fail to accurately categorize a scene from a remote sensing image recorded from the earth. A method that uses bilinear convolution neural networks to produce a lessweighted set of models those results in better visual recognition in remote sensing images using fine-grained techniques. This proposed hybrid method is utilized to extract scene feature information in two times from remote sensing images for improved recognition. In layman's terms, these features are defined as raw, and only have a single defined frame, so they will allow basic recognition from remote sensing images. This research work has proposed a double feature extraction hybrid deep learning approach to classify remotely sensed image scenes based on feature abstraction techniques. Also, the proposed algorithm is applied to feature values in order to convert them to feature vectors that have pure black and white values after many product operations. The next stage is pooling and normalization, which occurs after the CNN feature extraction process has changed. This research work has developed a novel hybrid framework method that has a better level of accuracy and recognition rate than any prior model.