High Spatial Resolution Remote Sensing Research Articles

Water Body Extraction from Very High Spatial Resolution Remote Sensing Data Based on Fully Convolutional Networks

This paper studies the use of the Fully Convolutional Networks (FCN) model in the extraction of water bodies from Very High spatial Resolution (VHR) optical images in the case of limited training samples. Two different seasonal GaoFen-2 images with a spatial resolution of 0.8 m in the south of the Beijing metropolitan area were used to extensively validate the FCN model. Four key factors including input features, training data, transfer learning, and data augmentation related to the performance of the FCN model were empirically analyzed by using 36 combinations of various parameter settings. Our findings indicate that the FCN-based method can work as a robust and cost-effective tool in the extraction of water bodies from VHR images. The FCN-based method trained on a small amount of labeled L1A data can also significantly outperform the Normalized Difference Water Index (NDWI) based method, the Support Vector Machine (SVM) based method, and the Sparsity Model (SM) based method, even when radiometric normalization and spatial contexts are introduced to preprocess the input data for the latter three methods. The advantages of the FCN-based method are mainly due to its capability to exploit spatial contexts in the image, especially in urban areas with mixed water and shadows. Though the settings of four key factors significantly affect the performance of the FCN based method, choosing a qualified setting for the FCN model is not difficult. Our lessons learned from the successful use of the FCN model for the extraction of water from VHR images can be extended to extract other land covers.

High Spatial Resolution Remote Sensing for Salt Marsh Mapping and Change Analysis at Fire Island National Seashore

Salt marshes are changing due to natural and anthropogenic stressors such as sea level rise, nutrient enrichment, herbivory, storm surge, and coastal development. This study analyzes salt marsh change at Fire Island National Seashore (FIIS), a nationally protected area, using object-based image analysis (OBIA) to classify a combination of data from Worldview-2 and Worldview-3 satellites, topobathymetric Light Detection and Ranging (LiDAR), and National Agricultural Imagery Program (NAIP) aerial imageries acquired from 1994 to 2017. The salt marsh classification was trained and tested with vegetation plot data. In October 2012, Hurricane Sandy caused extensive overwash and breached a section of the island. This study quantified the continuing effects of the breach on the surrounding salt marsh. The tidal inundation at the time of image acquisition was analyzed using a topobathymetric LiDAR-derived Digital Elevation Model (DEM) to create a bathtub model at the target tidal stage. The study revealed geospatial distribution and rates of change within the salt marsh interior and the salt marsh edge. The Worldview-2/Worldview-3 imagery classification was able to classify the salt marsh environments accurately and achieved an overall accuracy of 92.75%. Following the breach caused by Hurricane Sandy, bayside salt marsh edge was found to be eroding more rapidly (F1, 1597 = 206.06, p < 0.001). However, the interior panne/pool expansion rates were not affected by the breach. The salt marsh pannes and pools were more likely to revegetate if they had a hydrological connection to a mosquito ditch (χ2 = 28.049, p < 0.001). The study confirmed that the NAIP data were adequate for determining rates of salt marsh change with high accuracy. The cost and revisit time of NAIP imagery creates an ideal open data source for high spatial resolution monitoring and change analysis of salt marsh environments.

Instance Segmentation of Outdoor Sports Ground from High Spatial Resolution Remote Sensing Imagery Using the Improved Mask R-CNN

Instance Segmentation of Outdoor Sports Ground from High Spatial Resolution Remote Sensing Imagery Using the Improved Mask R-CNN

Spear and Shield: Attack and Detection for CNN-Based High Spatial Resolution Remote Sensing Images Identification

High spatial resolution remote sensing (HSRRS) images classification and identification is an important technology to acquire land surface information for land resource management, geographical situation monitoring, and global climate change. As the hottest deep learning method, convolutional neural network (CNN) has been successfully applied in HSRRS image classification and identification due to its powerful information extraction capability. However, adversarial perturbations caused by radiation transfer process or artificial or other unpredictable disturbances often deteriorate the stability of CNN. Under this background, we propose a robust architecture for adversarial attack and detection to classify and identify HSRRS images. First of all, two white-box attacks [i.e., large Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) and fast gradient sign method (FGSM)] are adopted respectively to generate adversarial images to confuse the model, and to assess the robustness of the HSRRS image classifier. Second, adversarial detection models based on support vector machine (SVM) with single or fused two level features are proposed to improve the detection accuracy. The features extracted from the testing CNN full connected layers contain adversarial perturbations and real information, from which SVM classifier and discriminate the real and the adversarial images. The adversarial attack model is evaluated in terms of overall accuracy (OA) and kappa coefficient (kc). The simulation results show that the OA decreases from 96.4% to 44.4% and 33.3% for L-BFGS and FGSM attacked classifier model, respectively. The adversarial detection is evaluated via OA, detection probability P D , false alarm probability P FA , and miss probability P M . The simulation results indicate that the fused model with two different level features based on SVM can obtain the best OA (94.5%), P D (0.933), P FA (0.040), and P M (0.067) among the detectors if the classifier is attacked by the FGSM. Meanwhile, when facing the L-BFGS attack, the fused model presents similar performance if the best single level features are utilized.

Deep Learning-Based Classification Methods for Remote Sensing Images in Urban Built-Up Areas

Urban areas have been focused recently on the remote sensing applications since their function closely relates to the distribution of built-up areas, where reflectivity or scattering characteristics are the same or similar. Traditional pixel-based methods cannot discriminate the types of urban built-up areas very well. This paper investigates a deep learning-based classification method for remote sensing images, particularly for high spatial resolution remote sensing (HSRRS) images with various changes and multi-scene classes. Specifically, to help develop the corresponding classification methods in urban built-up areas, we consider four deep neural networks (DNNs): 1) convolutional neural network (CNN); 2) capsule networks (CapsNet); 3) same model with a different training rounding based on CNN (SMDTR-CNN); and 4) same model with different training rounding based on CapsNet (SMDTR-CapsNet). The performances of the proposed methods are evaluated in terms of overall accuracy, kappa coefficient, precision, and confusion matrix. The results revealed that SMDTR-CNN obtained the best overall accuracy (95.0%) and kappa coefficient (0.944) while also improving the precision of parking lot and resident samples by 1% and 4%, respectively.

Mapping Urban Functional Zones by Integrating Very High Spatial Resolution Remote Sensing Imagery and Points of Interest: A Case Study of Xiamen, China

Fine-scale, accurate intra-urban functional zones (urban land use) are important for applications that rely on exploring urban dynamic and complexity. However, current methods of mapping functional zones in built-up areas with high spatial resolution remote sensing images are incomplete due to a lack of social attributes. To address this issue, this paper explores a novel approach to mapping urban functional zones by integrating points of interest (POIs) with social properties and very high spatial resolution remote sensing imagery with natural attributes, and classifying urban function as residence zones, transportation zones, convenience shops, shopping centers, factory zones, companies, and public service zones. First, non-built and built-up areas were classified using high spatial resolution remote sensing images. Second, the built-up areas were segmented using an object-based approach by utilizing building rooftop characteristics (reflectance and shapes). At the same time, the functional POIs of the segments were identified to determine the functional attributes of the segmented polygon. Third, the functional values—the mean priority of the functions in a road-based parcel—were calculated by functional segments and segmental weight coefficients. This method was demonstrated on Xiamen Island, China with an overall accuracy of 78.47% and with a kappa coefficient of 74.52%. The proposed approach could be easily applied in other parts of the world where social data and high spatial resolution imagery are available and improve accuracy when automatically mapping urban functional zones using remote sensing imagery. It will also potentially provide large-scale land-use information.

Detecting Building Edges from High Spatial Resolution Remote Sensing Imagery Using Richer Convolution Features Network

As the basic feature of building, building edges play an important role in many fields such as urbanization monitoring, city planning, surveying and mapping. Building edges detection from high spatial resolution remote sensing (HSRRS) imagery has always been a long-standing problem. Inspired by the recent success of deep-learning-based edge detection, a building edge detection model using a richer convolutional features (RCF) network is employed in this paper to detect building edges. Firstly, a dataset for building edges detection is constructed by the proposed most peripheral constraint conversion algorithm. Then, based on this dataset the RCF network is retrained. Finally, the edge probability map is obtained by RCF-building model, and this paper involves a geomorphological concept to refine edge probability map according to geometric morphological analysis of topographic surface. The experimental results suggest that RCF-building model can detect building edges accurately and completely, and that this model has an edge detection F-measure that is at least 5% higher than that of other three typical building extraction methods. In addition, the ablation experiment result proves that using the most peripheral constraint conversion algorithm can generate more superior dataset, and the involved refinement algorithm shows a higher F-measure and better visual effect contrasted with the non-maximal suppression algorithm.

Unsupervised Change Detection Based on Hybrid Conditional Random Field Model for High Spatial Resolution Remote Sensing Imagery

High spatial resolution (HSR) remote sensing images provide detailed geometric information about land cover. As a result, it is possible to detect more subtle changes with the help of HSR images. However, due to the increased spatial resolution and the limited spectral information, it is difficult to identify the real changes only through the spectral feature of the image. To fully explore the spectral–spatial information and improve the change detection performance for HSR images, this paper proposes the hybrid conditional random field (HCRF) model, which combines the traditional random field method with an object-based technique. In the proposed method, the spectral discriminative information of a single pixel is extracted by the unary potential, which is modeled using a soft clustering method to make an initial separation of changed and unchanged pixels. The pairwise potential then considers the contextual information of adjacent pixels to favor spatial smoothing. An object term is also introduced in the HCRF model to keep the homogeneity of changed objects. By the use of these approaches, the oversmoothing problem of the random field-based methods and the detection error caused by the segmentation strategy in the object-based methods can be relieved. The proposed method was tested on three HSR image data sets and outperformed the compared state-of-the-art techniques.

Remote Sensing Big Data: Theory, Methods and Applications

Nowadays, our ability to acquire remote sensing data has been improved to an unprecedented level.[...]

Road Extraction From a High Spatial Resolution Remote Sensing Image Based on Richer Convolutional Features

The extraction and vectorization of roads from high spatial resolution remote sensing (HSRRS) images are of great significance to city planning and development. However, significant as they are, it is usually an arduous task to put them into practice because the HSRRS images are often filled with complex ground information. Furthermore, extracted roads may suffer from netsplit or brokenness. This paper thus proposes Richer convolutional features (RCFs)-based road extraction (Road-RCF) as a method which targets these issues. A modified roads sample set and RCF network are applied to generate road probabilities in order to extract initial road information. After the road centerlines extraction by the refinement algorithm, vectorized roads are ultimately extracted. The compared experiment results show that the Road-RCF method produce better road extraction results than the other four state-of-the-art methods, in both quantitative road extraction accuracy metrics and the qualitative visual evaluation. The benefits of this model are threefold. First, the image-to-image network structure of side-output realizes multi-scale and multi-level road feature fusion in order to make a full use of the information from a low level to a high level. Second, according to the deep supervision of the side-output, it guides the learning of the correct road information. Third, after the detection of the road, the road centerlines are vectorized to facilitate the attribute information management and electronic map production. In a word, the proposed Road-RCF method is both practical and meaningful toward updating the geo-information system database.

A Sparse Dictionary Learning-Based Adaptive Patch Inpainting Method for Thick Clouds Removal from High-Spatial Resolution Remote Sensing Imagery.

Cloud cover is inevitable in optical remote sensing (RS) imagery on account of the influence of observation conditions, which limits the availability of RS data. Therefore, it is of great significance to be able to reconstruct the cloud-contaminated ground information. This paper presents a sparse dictionary learning-based image inpainting method for adaptively recovering the missing information corrupted by thick clouds patch-by-patch. A feature dictionary was learned from exemplars in the cloud-free regions, which was later utilized to infer the missing patches via sparse representation. To maintain the coherence of structures, structure sparsity was brought in to encourage first filling-in of missing patches on image structures. The optimization model of patch inpainting was formulated under the adaptive neighborhood-consistency constraint, which was solved by a modified orthogonal matching pursuit (OMP) algorithm. In light of these ideas, the thick-cloud removal scheme was designed and applied to images with simulated and true clouds. Comparisons and experiments show that our method can not only keep structures and textures consistent with the surrounding ground information, but also yield rare smoothing effect and block effect, which is more suitable for the removal of clouds from high-spatial resolution RS imagery with salient structures and abundant textured features.

Pre-Trained AlexNet Architecture with Pyramid Pooling and Supervision for High Spatial Resolution Remote Sensing Image Scene Classification

The rapid development of high spatial resolution (HSR) remote sensing imagery techniques not only provide a considerable amount of datasets for scene classification tasks but also request an appropriate scene classification choice when facing with finite labeled samples. AlexNet, as a relatively simple convolutional neural network (CNN) architecture, has obtained great success in scene classification tasks and has been proven to be an excellent foundational hierarchical and automatic scene classification technique. However, current HSR remote sensing imagery scene classification datasets always have the characteristics of small quantities and simple categories, where the limited annotated labeling samples easily cause non-convergence. For HSR remote sensing imagery, multi-scale information of the same scenes can represent the scene semantics to a certain extent but lacks an efficient fusion expression manner. Meanwhile, the current pre-trained AlexNet architecture lacks a kind of appropriate supervision for enhancing the performance of this model, which easily causes overfitting. In this paper, an improved pre-trained AlexNet architecture named pre-trained AlexNet-SPP-SS has been proposed, which incorporates the scale pooling—spatial pyramid pooling (SPP) and side supervision (SS) to improve the above two situations. Extensive experimental results conducted on the UC Merced dataset and the Google Image dataset of SIRI-WHU have demonstrated that the proposed pre-trained AlexNet-SPP-SS model is superior to the original AlexNet architecture as well as the traditional scene classification methods.

Classification of Urban Building Type from High Spatial Resolution Remote Sensing Imagery Using Extended MRS and Soft BP Network

This study presents a new approach for classification of building type in complex urban scene. The approach consists of two parts: extended multiresolution segmentation (EMRS) and soft classification using BP network (SBP). The technology scheme is referred to here as EMRS-SBP. EMRS is used to guide the design of descriptor. A descriptor is a feature expression or a symbolized algorithm to systematically promote the expressing capability of image features. A classifier can perform far better to discern complex pattern of combining pixels working in an EMRS-based feature space constructed by a number of such descriptors. SBP serves as a classifier model to generate natural clusters of member which refers to here as both pixels and image patches. Class-mark ensured member is denoted as sure member and the rest as unsure (fuzzy) members. The latter can be relabeled through recursive defuzzifying according to the information carried by the gradually increased sure members. By using EMRS-SBP, three building types, i.e., old-fashioned courtyard dwellings, multistorey residential buildings, and high-rise buildings, can be accurately classified from high spatial resolution imagery in a feature space constructed with fifteen descriptors including nine EMRS-based ones. There is evidence that the mean overall accuracy using SBP in the EMRS-based feature space is 19.8% higher than that using the hard classification with BP network in a single resolution segmentation space and meanwhile, the mean kappa statistic value ( κ ) is 25.1% higher.

Zero-Shot Scene Classification for High Spatial Resolution Remote Sensing Images

Due to the rapid technological development of various sensors, a huge volume of high spatial resolution (HSR) image data can now be acquired. How to efficiently recognize the scenes from such HSR image data has become a critical task. Conventional approaches to remote sensing scene classification only utilize information from HSR images. Therefore, they always need a large amount of labeled data and cannot recognize the images from an unseen scene class without any visual sample in the labeled data. To overcome this drawback, we propose a novel approach for recognizing images from unseen scene classes, i.e., zero-shot scene classification (ZSSC). In this approach, we first use the well-known natural language process model, word2vec, to map names of seen/unseen scene classes to semantic vectors. A semantic-directed graph is then constructed over the semantic vectors for describing the relationships between unseen classes and seen classes. To transfer knowledge from the images in seen classes to those in unseen classes, we make an initial label prediction on test images by an unsupervised domain adaptation model. With the semantic-directed graph and initial prediction, a label-propagation algorithm is then developed for ZSSC. By leveraging the visual similarity among images from the same scene class, a label refinement approach based on sparse learning is used to suppress the noise in the zero-shot classification results. Experimental results show that the proposed approach significantly outperforms the state-of-the-art approaches in ZSSC.

An Efficient and Robust Integrated Geospatial Object Detection Framework for High Spatial Resolution Remote Sensing Imagery

Geospatial object detection from high spatial resolution (HSR) remote sensing imagery is a significant and challenging problem when further analyzing object-related information for civil and engineering applications. However, the computational efficiency and the separate region generation and localization steps are two big obstacles for the performance improvement of the traditional convolutional neural network (CNN)-based object detection methods. Although recent object detection methods based on CNN can extract features automatically, these methods still separate the feature extraction and detection stages, resulting in high time consumption and low efficiency. As a significant influencing factor, the acquisition of a large quantity of manually annotated samples for HSR remote sensing imagery objects requires expert experience, which is expensive and unreliable. Despite the progress made in natural image object detection fields, the complex object distribution makes it difficult to directly deal with the HSR remote sensing imagery object detection task. To solve the above problems, a highly efficient and robust integrated geospatial object detection framework based on faster region-based convolutional neural network (Faster R-CNN) is proposed in this paper. The proposed method realizes the integrated procedure by sharing features between the region proposal generation stage and the object detection stage. In addition, a pre-training mechanism is utilized to improve the efficiency of the multi-class geospatial object detection by transfer learning from the natural imagery domain to the HSR remote sensing imagery domain. Extensive experiments and comprehensive evaluations on a publicly available 10-class object detection dataset were conducted to evaluate the proposed method.

Seasonal Timing for Estimating Carbon Mitigation in Revegetation of Abandoned Agricultural Land with High Spatial Resolution Remote Sensing

Dryland salinity is a major land management issue globally, and results in the abandonment of farmland. Revegetation with halophytic shrub species such as Atriplex nummularia for carbon mitigation may be a viable option but to generate carbon credits ongoing monitoring and verification is required. This study investigated the utility of high-resolution airborne images (Digital Multi Spectral Imagery (DMSI)) obtained in two seasons to estimate carbon stocks at the plant- and stand-scale. Pixel-scale vegetation indices, sub-pixel fractional green vegetation cover for individual plants, and estimates of the fractional coverage of the grazing plants within entire plots, were extracted from the high-resolution images. Carbon stocks were correlated with both canopy coverage (R2: 0.76–0.89) and spectral-based vegetation indices (R2: 0.77–0.89) with or without the use of the near-infrared spectral band. Indices derived from the dry season image showed a stronger correlation with field measurements of carbon than those derived from the green season image. These results show that in semi-arid environments it is better to estimate saltbush biomass with remote sensing data in the dry season to exclude the effect of pasture, even without the refinement provided by a vegetation classification. The approach of using canopy cover to refine estimates of carbon yield has broader application in shrublands and woodlands.

Carbon Sequestration Analysis of dominant tree species using Geo-informatics Technology in Gujarat State (INDIA)

Comparative analysis of carbon store of three dominant tree species in planned green capital city of Gandhinagar and unplanned industrial town of Mahesana in Gujarat State, India was carried out using geo-informatics technology. The selected three major dominant tree species grown in these towns are Azadirachta indica, Acacia sp. and Cassia sp. The high spatial resolution Remote Sensing Satellite data from Indian Remote Sensing Satellite (IRS-Resourcesat-1) LISS-IV, Cartosat-1, and Google earth images were used in this study. The tree cover was categorized into dense and sparse on the basis of canopy cover observed on Satellite data. A grid of 1km X 1km was created in GIS environment and superimposed on Cartosat-1 images. Random sample of 20 % was selected for detailed tree count in the field and total tree count was estimated from these selected sample grids. Total biomass and carbon sequestered in the major tree species have been estimated using a non-destructive method. The carbon stock estimated for three major tree species in Gandhinagar and Mahesana towns indicate that Azadirachta indica has maximum carbon sequestration potential as compared to Acacia sp. and Cassia sp. The maximum of carbon stock was present in Girth at Breast Height (GBH) size >180 cm which is followed by GBH size 90 to 180 cm. The total number of trees in Gandhinagar town is much higher as compared to Mahesana town; therefore estimated carbon store of dominant tree species in Gandhinagar town is very high as compared to Mahesana town which is more arid as compared to Gandhinagar.

Assessing the Benefits from Retrofitting Green Roofs in Mediterranean, Using Environmental Modelling, GIS and Very High Spatial Resolution Remote Sensing Data: The Example of Thessaloniki, Greece

In the Mediterranean area, rural abandonment and the rapid urbanisation after the 1950's led to complex environmental problems, such as the exploitation of natural resources, the environmental pollution, whilst the unstrained urban growth caused excessive pressure on the existing infrastructure, which affects buildings, public transportation, water quality and public health.Various studies demonstrate that the implementation of green roof technology can moderate environmental problems, through the reduction of heat flux and solar reflectivity, the minimisation of buildings’ energy consumption, the air pollution removal, the storm water runoff reduction, the air cooling and the effective management of the urban heat island effect.In this framework, the aim of this study is to assess the green roof potential and the quantification of its benefits over Thessaloniki, Greece's second largest city. In order to do so, very high spatial resolution satellite image and natural colour orthoimagery were used along with a geographical object-based image analysis approach. Carbon sequestration potential, rainwater retention and energy conservation were estimated based on coefficients adopted after a comprehensive literature review and actual dynamic energy simulations.

Change Detection Based on a Multifeature Probabilistic Ensemble Conditional Random Field Model for High Spatial Resolution Remote Sensing Imagery

In this letter, a multifeature probabilistic ensemble conditional random field (MFPECRF) model is proposed to perform the task of change detection for high spatial resolution (HSR) remote sensing imagery. MFPECRF not only considers the spectral feature of single pixels but also the interaction between neighborhood pixels and the structural property of the ground objects in HSR imagery to give a higher detection accuracy than the traditional random field methods, which only utilize spectral and label information. In the unary potential, the spectral and morphological features of the difference image are combined using a probabilistic ensemble strategy, and the pairwise potential considers the contextual information of the observed field. The parameters of MFPECRF are estimated using a piecewise strategy, and the final result is obtained by the use of the loopy belief propagation algorithm. The experimental results of two groups of HSR multispectral images confirm the potential of the proposed method in improving the detection accuracy for HSR imagery.

SatCNN: satellite image dataset classification using agile convolutional neural networks

ABSTRACTWith the launch of various remote-sensing satellites, more and more high-spatial resolution remote-sensing (HSR-RS) images are becoming available. Scene classification of such a huge volume of HSR-RS images is a big challenge for the efficiency of the feature learning and model training. The deep convolutional neural network (CNN), a typical deep learning model, is an efficient end-to-end deep hierarchical feature learning model that can capture the intrinsic features of input HSR-RS images. However, most published CNN architectures are borrowed from natural scene classification with thousands of training samples, and they are not designed for HSR-RS images. In this paper, we propose an agile CNN architecture, named as SatCNN, for HSR-RS image scene classification. Based on recent improvements to modern CNN architectures, we use more efficient convolutional layers with smaller kernels to build an effective CNN architecture. Experiments on SAT data sets confirmed that SatCNN can quickly and effectively learn robust features to handle the intra-class diversity even with small convolutional kernels, and the deeper convolutional layers allow spontaneous modelling of the relative spatial relationships. With the help of fast graphics processing unit acceleration, SatCNN can be trained within about 40 min, achieving overall accuracies of 99.65% and 99.54%, which is the state-of-the-art for SAT data sets.