Remote Sensing Scene Classification Based on Convolutional Neural Networks Pre-Trained Using Attention-Guided Sparse Filters

Scene classification of high-resolution remote sensing images (HRRSI) is one of the most important means of land-cover classification. Deep learning techniques, especially the convolutional neural network (CNN) have been widely applied to the scene classification of HRRSI due to the advancement of graphic processing units (GPU). However, they tend to extract features from the whole images rather than discriminative regions. The visual attention mechanism can force the CNN to focus on discriminative regions, but it may suffer from the influence of intra-class diversity and repeated texture. Motivated by these problems, we propose an attention-based deep feature fusion (ADFF) framework that constitutes three parts, namely attention maps generated by Gradient-weighted Class Activation Mapping (Grad-CAM), a multiplicative fusion of deep features and the center-based cross-entropy loss function. First of all, we propose to make attention maps generated by Grad-CAM as an explicit input in order to force the network to concentrate on discriminative regions. Then, deep features derived from original images and attention maps are proposed to be fused by multiplicative fusion in order to consider both improved abilities to distinguish scenes of repeated texture and the salient regions. Finally, the center-based cross-entropy loss function that utilizes both the cross-entropy loss and center loss function is proposed to backpropagate fused features so as to reduce the effect of intra-class diversity on feature representations. The proposed ADFF architecture is tested on three benchmark datasets to show its performance in scene classification. The experiments confirm that the proposed method outperforms most competitive scene classification methods with an average overall accuracy of 94% under different training ratios.

Land-use scene classification (LUSC) is a key technique in the field of remote sensing imagery (RSI) interpretation. A convolutional neural network (CNN) is widely used for its ability to autonomously and efficiently extract deep semantic feature maps (DSFMs) from large-scale RSI data. However, CNNs cannot accurately extract the rich spatial structure information of RSI, and the key information of RSI is easily lost due to many pooling layers, so it is difficult to ensure the information integrity of the spatial structure feature maps (SSFMs) and DSFMs of RSI with CNNs only for LUSC, which can easily affect the classification performance. To fully utilize the SSFMs and make up for the insufficiency of CNN in capturing the relationship information between the land-use objects of RSI, while reducing the loss of important information, we propose an effective dual-branch hybrid framework, HFCC-Net, for the LUSC task. The CNN in the upper branch extracts multi-scale DSFMs of the same scene using transfer learning techniques; the graph routing-based CapsNet in the lower branch is used to obtain SSFMs from DSFMs in different scales, and element-by-element summation achieves enhanced representations of SSFMs; a newly designed function is used to fuse the top-level DSFMs with SSFMs to generate discriminant feature maps (DFMs); and, finally, the DFMs are fed into classifier. We conducted sufficient experiments using HFCC-Net on four public datasets. The results show that our method has better classification performance compared to some existing CNN-based state-of-the-art methods.

Prediction of PM2.5 concentration in urban agglomeration of China by hybrid network model

In view of the scene’s complexity and diversity in scene classification, this paper makes full use of the contextual semantic relationships between the objects to describe the visual attention regions of the scenes and combines with the deep convolution neural networks, so that a scene classification model using visual attention and deep networks is constructed. Firstly, the visual attention regions in the scene image are marked by using the context-based saliency detection algorithm. Then, the original image and the visual attention region detection image are superimposed to obtain a visual attention region enhancement image. Furthermore, the deep convolution features of the original image, the visual attention region detection image, and the visual attention region enhancement image are extracted by using the deep convolution neural networks pretrained on the large-scale scene image dataset Places. Finally, the deep visual attention features are constructed by using the multilayer deep convolution features of the deep convolution networks, and a classification model is constructed. In order to verify the effectiveness of the proposed model, the experiments are carried out on four standard scene datasets LabelMe, UIUC-Sports, Scene-15, and MIT67. The results show that the proposed model improves the performance of the classification well and has good adaptability.

Remote sensing (RS) scene classification plays an important role in the field of earth observation. With the rapid development of the RS techniques, a large number of RS scene images are available. As manually labeling large-scale RS scene images is both labor and time consuming, when a new unlabeled data set is obtained, how to use the existing labeled data sets to classify the new unlabeled images is an important research direction. Different RS scene image data sets may be taken from different type of sensors, and the images may vary from imaging modalities, spatial resolutions, and image scales, so the distribution discrepancy exists among different image data sets. As a result, simply applying convolutional neural networks (CNN) trained on source domain cannot accurately classify the images on target domain. Domain adaptation (DA) can be helpful to solve this problem. In this letter, we design a subspace alignment (SA) and CNN-based framework to solve the DA problem in RS scene image classification. A new SA layer is proposed and added into CNN models for DA, which could align the source and target domains in some feature subspace. Fine-tuning the modified CNN model with the added SA layer makes the CNN model adapt to the aligned feature subspace and helps to relieve the domain distribution discrepancy. The experiments conducted on two public data sets show that adding the SA layer into CNN improves the scene classification on the target domain.

Acoustic scene classification is the task of identifying the scene from which the audio signal is recorded. Convolutional neural network (CNN) models are widely adopted with proven successes in acoustic scene classification. However, there is little insight on how an audio scene is perceived in CNN, as what have been demonstrated in image recognition research. In the present study, the Class Activation Mapping (CAM) is utilized to analyze how the log-magnitude Mel-scale filter-bank (log-Mel) features of different acoustic scenes are learned in a CNN classifier. It is noted that distinct high-energy time-frequency components of audio signals generally do not correspond to strong activation on CAM, while the background sound texture are well learned in CNN. In order to make the sound texture more salient, we propose to apply the Difference of Gaussian (DoG) and Sobel operator to process the log-Mel features and enhance edge information of the time-frequency image. Experimental results on the DCASE 2017 ASC challenge show that using edge enhanced log-Mel images as input feature of CNN significantly improves the performance of audio scene classification.

paper, we focus on exploring the relationship between weave patterns and their mechanical properties in woven fiber composites through Machine Learning. Specifically, we explore the interactions between woven architectures and in-plane stiffness properties through Deep Convolutional Neural Network (DCNN) and Generative Adversarial Network (GAN). Our research is important for exploring how woven composite’s pattern is related to its mechanical properties and accelerating woven composite design as well as optimization. We focus on two tasks: (1) Stiffness prediction: Predicting in-plane stiffness properties for given weave patterns. Our DCNN extracts high-level features through several convolutional and fully connected layers to determine the final predictions. (2) Weave pattern prediction: Predicting weave patterns for target stiffness properties, which can be treated as the reverse task of the first one. Due to many-to-one mapping between weave patterns and the composite properties, we utilize a Decoder Neural Network as our baseline model and compare its performance with GAN and Genetic Algorithm. We represent the weave patterns as 2D checkerboard models and use finite element analysis (FEA) to determine in-plane stiffness properties, which serve as input data for our ML framework. We show that: (1) for stiffness prediction, DCNN can predict stiffness values for a given weave pattern with relatively high accuracy (above 93%); (2) for weave pattern prediction, the GAN model gives the best prediction accuracy (above 92%) while Decoder Neural Network has the best time efficiency. HAOTIAN FENG

Scene classification of high spatial resolution (HSR) images can provide data support for many practical applications, such as land planning and utilization, and it has been a crucial research topic in the remote sensing (RS) community. Recently, deep learning methods driven by massive data show the impressive ability of feature learning in the field of HSR scene classification, especially convolutional neural networks (CNNs). Although traditional CNNs achieve good classification results, it is difficult for them to effectively capture potential context relationships. The graphs have powerful capacity to represent the relevance of data, and graph-based deep learning methods can spontaneously learn intrinsic attributes contained in RS images. Inspired by the abovementioned facts, we develop a deep feature aggregation framework driven by graph convolutional network (DFAGCN) for the HSR scene classification. First, the off-the-shelf CNN pretrained on ImageNet is employed to obtain multilayer features. Second, a graph convolutional network-based model is introduced to effectively reveal patch-to-patch correlations of convolutional feature maps, and more refined features can be harvested. Finally, a weighted concatenation method is adopted to integrate multiple features (i.e., multilayer convolutional features and fully connected features) by introducing three weighting coefficients, and then a linear classifier is employed to predict semantic classes of query images. Experimental results performed on the UCM, AID, RSSCN7, and NWPU-RESISC45 data sets demonstrate that the proposed DFAGCN framework obtains more competitive performance than some state-of-the-art methods of scene classification in terms of OAs.

Scene change detection is the process of identifying the differences between the multi-temporal image scenes at the semantic level, which has significant potential in the application of urban development and land management. In this paper, we propose a novel deep convolution canonical correlation analysis neural network (DCCANet) architecture, which could consider the spectral-spatial-temporal correlation feature for scene change detection in remote sensing images. For this purpose, we put together the convolutional neural networks (CNNs) and the deep canonical correlation analysis (DCCA) into the end-to-end network. The CNN could get the spectral-spatial feature information for scene representation, while the later could enhance the temporal correlation by nonlinear high- dimensional transformations between the multi-temporal image scenes for scene change detection. Experiments with high-resolution remote sensing image scene datasets demonstrated that our proposed approach can get a better performance in scene classification and change detection.

Seed freezing damage is an agricultural disaster, which has a great impact on seed quality assurance. The feasibility of combining hyperspectral imaging with deep convolutional neural network (DCNN) to classify different freeze-damaged corn seeds was studied in this paper. At first, the hyperspectral images of corn seeds subjected to five different freezing temperatures at 400–1000 nm were acquired, and then the average spectra were extracted from the region of embryo hyperspectral images over the wavelength range of 450–979 nm. Next, four models (K nearest neighbors (KNN), support vector machine (SVM), extreme learning machine (ELM), and DCNN) were developed for five-category (5 frozen conditions) and four-category (“no freezing,” “slight freezing,” “moderate freezing,” and “severe freezing”) classifications, and the values of the evaluation indexes (accuracy, sensitivity, specificity, and precision) were calculated for comparison. The results show that DCNN model had the most satisfactory result with accuracy rates of 100% (training set), 96.9% (validation set), and 97.5% (testing set) for five-category classification, with accuracy rates of 100% (training, validation, and testing set) for four-category classification, and DCNN model also had the best performance in the evaluation indexes. At last, the visual classification map was generated according to the results of DCNN. It shows that hyperspectral imaging and DCNN can provide a novel method to detect the freezing damage of corn seeds quickly and inexpensively.

ABSTRACTIn recent years, object-based segmentation methods and shallow-model classification algorithms have been widely integrated for remote sensing image supervised classification. However, as the image resolution increases, remote sensing images contain increasingly complex characteristics, leading to higher intraclass heterogeneity and interclass homogeneity and thus posing substantial challenges for the application of segmentation methods and shallow-model classification algorithms. As important methods of deep learning technology, convolutional neural networks (CNNs) can hierarchically extract higher-level spatial features from images, providing CNNs with a more powerful recognition ability for target detection and scene classification in high-resolution remote sensing images. However, the input of the traditional CNN is an image patch, the shape of which is scarcely consistent with a given segment. This inconsistency may lead to errors when directly using CNNs in object-based remote sensing classification: jagged errors may appear along the land cover boundaries, and some land cover areas may overexpand or shrink, leading to many obvious classification errors in the resulting image. To address the above problem, this paper proposes an object-based and heterogeneous segment filter convolutional neural network (OHSF-CNN) for high-resolution remote sensing image classification. Before the CNN processes an image patch, the OHSF-CNN includes a heterogeneous segment filter (HSF) to process the input image. For the segments in the image patch that are obviously different from the segment to be classified, the HSF can differentiate them and reduce their negative influence on the CNN training and decision-making processes. Experimental results show that the OHSF-CNN not only can take full advantage of the recognition capabilities of deep learning methods but also can effectively avoid the jagged errors along land cover boundaries and the expansion/shrinkage of land cover areas originating from traditional CNN structures. Moreover, compared with the traditional methods, the proposed OHSF-CNN can achieve higher classification accuracy. Furthermore, the OHSF-CNN algorithm can serve as a bridge between deep learning technology and object-based segmentation algorithms thereby enabling the application of object-based segmentation methods to more complex high-resolution remote sensing images.

Remote sensing image (RSI) scene classification is the foundation and important technology of ground object detection, land use management and geographic analysis. During recent years, convolutional neural networks (CNNs) have achieved significant success and are widely applied in RSI scene classification. However, crafted images that serve as adversarial examples can potentially fool CNNs with high confidence and are hard for human eyes to interpret. For the increasing security and robust requirements of RSI scene classification, the adversarial example problem poses a serious problem for the classification results derived from systems using CNN models, which has not been fully recognized by previous research. In this study, to explore the properties of adversarial examples of RSI scene classification, we create different scenarios by testing two major attack algorithms (i.e., the fast gradient sign method (FGSM) and basic iterative method (BIM)) trained on different RSI benchmark datasets to fool CNNs (i.e., InceptionV1, ResNet and a simple CNN). In the experiment, our results show that CNNs of RSI scene classification are also vulnerable to adversarial examples, and some of them have a fooling rate of over 80%. These adversarial examples are affected by the architecture of CNNs and the type of RSI dataset. InceptionV1 has a fooling rate of less than 5%, which is lower than the others. Adversarial examples generated on the UCM dataset are easier than other datasets. Importantly, we also find that the classes of adversarial examples have an attack selectivity property. Misclassifications of adversarial examples of RSIs are related to the similarity of the original classes in the CNN feature space. Attack selectivity reveals potential classes of adversarial examples and provides insights into the design of defensive algorithms in future research.

In this paper, we explore learning methods to improve the performance of the open-circuit fault diagnosis of modular multilevel converters (MMCs). Two deep learning methods, namely, convolutional neural networks (CNN) and auto encoder based deep neural networks (AE-based DNN), as well as stand-alone SoftMax classifier are explored for the detection and classification of faults of MMC-based high voltage direct current converter (MMC-HVDC). Only AC-side three-phase current and the upper and lower bridges’ currents of the MMCs are used directly in our proposed approaches without any explicit feature extraction or feature subset selection. The two-terminal MMC-HVDC system is implemented in Power Systems Computer-Aided Design/Electromagnetic Transients including DC (PSCAD/EMTDC) to verify and compare our methods. The simulation results indicate CNN, AE-based DNN, and SoftMax classifier can detect and classify faults with high detection accuracy and classification accuracy. Compared with CNN and AE-based DNN, the SoftMax classifier performed better in detection and classification accuracy as well as testing speed. The detection accuracy of AE-based DNN is a little better than CNN, while CNN needs less training time than the AE-based DNN and SoftMax classifier.

Multimodal alias, guided, image super-resolution (SR) refers to the reconstruction of a high-resolution (HR) version of a low-resolution (LR) image with the aid of an HR image from another image modality. Common approaches for the SR problem include analytical methods which are computationally expensive. Deep learning methods are capable of learning a nonlinear mapping between LR and HR images from data, delivering high reconstruction accuracy at a low-computational cost during inference; however, these methods do not incorporate any prior knowledge about the problem, with the neural network model behaving like a black box. In this paper, we formulate multimodal image SR as a coupled convolutional sparse coding problem. To solve the corresponding minimization problem, we adopt the Method of Multipliers (MM). We then design a convolutional neural network (CNN) that unfolds the obtained MM algorithm. The proposed CNN accepts as input the LR image from the main modality and the HR image from the guidance modality to reconstruct the desired HR image. Unlike existing deep learning methods, our CNN provides an efficient and structured way to fuse information at different stages of the network and achieves high reconstruction accuracy. We evaluate the performance of the proposed model for the super-resolution of multi-spectral images guided by their high resolution RGB counterparts.

While convolutional neural networks (CNN) have been excellent for object recognition, the greater spatial variability in scene images typically meant that the standard full-image CNN features are suboptimal for scene classification. In this paper, we investigate a framework allowing greater spatial flexibility, in which the Fisher vector (FV) encoded distribution of local CNN features, obtained from a multitude of region proposals per image, is considered instead. The CNN features are computed from an augmented pixel-wise representation comprising multiple modalities of RGB, HHA and surface normals, as extracted from RGB-D data. More significantly, we make two postulates: (1) component sparsity - that only a small variety of region proposals and their corresponding FV GMM components contribute to scene discriminability, and (2) modal non-sparsity - within these discriminative components, all modalities have important contribution. In our framework, these are implemented through regularization terms applying group lasso to GMM components and exclusive group lasso across modalities. By learning and combining regressors for both proposal-based FV features and global CNN features, we were able to achieve state-of-the-art scene classification performance on the SUNRGBD Dataset and NYU Depth Dataset V2.