Change Detection Model Research Articles

Change detection with various combinations of fluid pyramid integration networks

An increasing number of change detection models are designed with different convolutional neural network (CNNs). However, the mechanism for designing network layers that can effectively extract robust features for different scenes remains unclear. Thus, novel networks with fluid pyramid integration network (FPIN) to detect changes are proposed in this study. Specifically, we first extract multi-scale deep features from feature extraction network. Higher layers extract semantic features robust to illumination and camera pose variations, while lower layers extract texture features that generate clear boundaries and changes with small scales. To employ the advantages of the features extracted from different layers, FPIN progressively fuses higher and lower layer features in a hierarchical order. We propose three different change detection networks based on FPIN and validate them on three publicly available change detection benchmark datasets. Experimental results showed that the proposed networks are better than state-of-the-art change detection methods.

Topographically derived subpixel-based change detection for monitoring changes over rugged terrain Himalayas using AWiFS data

Continuous and accurate monitoring of earth surface changes over rugged terrain Himalayas is important to manage natural resources and mitigate natural hazards. Conventional techniques generally focus on per-pixel based processing and overlook the sub-pixel variations occurring especially in case of low or moderate resolution remotely sensed data. However, the existing subpixel-based change detection (SCD) models are less effective to detect the mixed pixel information at its complexity level especially over rugged terrain regions. To overcome such issues, a topographically controlled SCD model has been proposed which is an improved version of widely used per-pixel based change vector analysis (CVA) and hence, named as a subpixel-based change vector analysis (SCVA). This study has been conducted over a part of the Western Himalayas using the advanced wide-field sensor (AWiFS) and Landsat-8 datasets. To check the effectiveness of the proposed SCVA, the cross-validation of the results has been done with the existing neural network-based SCD (NN-SCD) and per-pixel based models such as fuzzy-based CVA (FCVA) and post-classification comparison (PCC). The results have shown that SCVA offered robust performance (85.6%–86.4%) as compared to NN-SCD (81.6%–82.4%), PCC (79.2%–80.4%), and FCVA (81.2%–83.6%). We concluded that SCVA helps in reducing the detection of spurious pixels and improve the efficacy of generating change maps. This study is beneficial for the accurate monitoring of glacier retreat and snow cover variability over rugged terrain regions using moderate resolution remotely sensed datasets.

Research on Change Detection Method of High-Resolution Remote Sensing Images Based on Subpixel Convolution

Remote sensing image change detection method plays a great role in land cover research, disaster assessment, medical diagnosis, video surveillance, and other fields, so it has attracted wide attention. Based on a small sample dataset from SZTAKI AirChange Benchmark Set, in order to solve the problem that the deep learning network needs a large number of samples, this work first uses nongenerative sample augmentation method and generative sample augmentation method based on deep convolutional generative adversarial networks, and then, constructs a remote sensing image change detection model based on an improved DeepLabv3+ network. This model can realize end-to-end training and prediction of remote sensing image change detection with subpixel convolution. Finally, Landsat 8, Google Earth, and Onera satellite change detection datasets are used to verify the generalization performance of this network. The experimental results show that the improved network accuracy is 95.1% and the generalization performance is acceptable.

Soil Moisture Estimation based on the Distributed Scatterers Adaptive Filter over the QTP Permafrost Region using Sentinel-1 and High-resolution TerraSAR-X Data

ABSTRACT Synthetic aperture radar (SAR) data have been widely and primarily used for soil-moisture estimation. The original radar backscattering coefficients show great noise and temporal-spatial decorrelation due to complex surface features and seasonal vegetation changes, which increase the difficulty of soil-moisture estimation over the Qinghai-Tibet Plateau (QTP). To overcome these limitations, a soil-moisture retrieval method based on the Distributed Scatterers Adaptive Filter (DSAF) is proposed and applied in this paper. First, the statistically homogeneous pixels (SHP) with similar scatter behaviour are identified using the Anderson-Darling (AD) test method. Then, the Distributed Scatterers (DSs) are selected according to the SHPs threshold, and the backscattering coefficients of DSs are filtered using the SHPs distribution and coherences. In order to eliminate the negative influence of vegetation and roughness, the filtered backscattering coefficients are further processed by coupling the Water Cloud Model (WCM) and the images acquired in the winter season. The response of soil moisture to backscattering coefficients is greatly improved. Finally, the backscatter coefficients () are transformed into soil-moisture units by using the Least Square Regression (LSR) model and multi-temporal Change Detection (CD) method, respectively. The 15 scenes of TerraSAR-X images and 18 scenes of Sentinel-1 images in Northern Piedmont river, Northern Tibet, were used for experimental study. The results reveal that the response of the filtered backscattering coefficients to in-situ soil moisture is greatly improved. The soil-moisture content estimated by LSR and CD model using TerraSAR-X data is consistent with in-situ measurements, with coefficient of determination (R 2) of 0.52 and 0.56, Root Mean Square Error (RMSE) of 0.08 and 0.06, respectively. The results show that the proposed method can improve the accuracy of soil-moisture estimation, especially over the complex surface features of the QTP permafrost region.

3DCD: Scene Independent End-to-End Spatiotemporal Feature Learning Framework for Change Detection in Unseen Videos.

Change detection is an elementary task in computer vision and video processing applications. Recently, a number of supervised methods based on convolutional neural networks have reported high performance over the benchmark dataset. However, their success depends upon the availability of certain proportions of annotated frames from test video during training. Thus, their performance on completely unseen videos or scene independent setup is undocumented in the literature. In this work, we present a scene independent evaluation (SIE) framework to test the supervised methods in completely unseen videos to obtain generalized models for change detection. In addition, a scene dependent evaluation (SDE) is also performed to document the comparative analysis with the existing approaches. We propose a fast (speed-25 fps) and lightweight (0.13 million parameters, model size-1.16 MB) end-to-end 3D-CNN based change detection network (3DCD) with multiple spatiotemporal learning blocks. The proposed 3DCD consists of a gradual reductionist block for background estimation from past temporal history. It also enables motion saliency estimation, multi-schematic feature encoding-decoding, and finally foreground segmentation through several modular blocks. The proposed 3DCD outperforms the existing state-of-the-art approaches evaluated in both SIE and SDE setup over the benchmark CDnet 2014, LASIESTA and SBMI2015 datasets. To the best of our knowledge, this is a first attempt to present results in clearly defined SDE and SIE setups in three change detection datasets.

Accelerate disaster risk mapping using an automated zonal change detection model for loss estimation and recovery: a flood case study

BackgroundThe flash flood still constitutes one of the major natural meteorological disasters harmfully threatening local communities, that creates life losses and destroying infrastructures. The severity and magnitude of disasters always reflected from the size of impacts. Most of the conventional research models related to flooding vulnerability are focusing on hydro-meteorological and morphometric measurements. It, however, requires quick estimate of the flood losses and assess the severity using reliable information. An automated zonal change detection model applied, using two high-resolution satellite images dated 2009 and 2011 coupled with LU/LC GIS layer, on western El-Arish City, downstream of Wadi El-Arish basin. The model enabled to estimate the severity of a past flood incident in 2010.ResultsThe model calculated the total changes based on the before and after satellite images based on pixel-by-pixel comparison. The estimated direct-damages nearly 32,951 m2 of the total mapped LU/LC classes; (e.g., 11,407 m2 as 3.17% of the cultivated lands; 6031 m2 as 7.22% of the built-up areas and 4040 m2 as 3.62% of the paved roads network). The estimated cost of losses, in 2010 economic prices for the selected three LU/LC classes, is nearly 25 million USD, for the cultivation fruits and olives trees, ~ 4 million USD for built-up areas and ~ 1 million USD for paved roads network.ConclusionThe disasters’ damage and loss estimation process takes many detailed data, longtime, and costed as well. The applied model accelerates the disaster risk mapping that provides an informative support for loss estimation. Therefore, decision-makers and professionals need to apply this model for quick the disaster risks management and recovery.

Optical Remote Sensing Image Change Detection Based on Attention Mechanism and Image Difference

This study presents a new end-to-end change detection network, called difference-enhancement dense-attention convolutional neural network (DDCNN), that is designed for detection of changes in the bitemporal optical remote sensing images. To model the internal correlation between high-level and low-level features, a dense attention method consisting of several up-sampling attention units is proposed. Both the up-sampling spatial and up-sampling channel attention are adopted by the unit. The unit, which can use high-level features with rich category information to guide the selection of low-level features, can use the spatial context information to capture the changed features of ground objects. Furthermore, DDCNN also pays attention to the differentiating features of the bitemporal images. By introducing a DE unit, each pixel is weighted and the features are selectively aggregated. The combination of dense attention and the DE unit improves the effectiveness of the network and its accuracy in extracting the change features. The effectiveness of the proposed approach is demonstrated via five challenge data sets. The experimental results show that DDCNN achieves new state-of-the-art change detection performance on these five challenging data sets. For the seasonal change detection data set in particular, compared with the best existing change detection model, the proposed method increases the F1 score and IoU by 2.96% and 5.17%, respectively; compared with the baseline method, our method improved 3.75% and 6.50% on the F1 score and IoU, respectively.

Deep Learning for Regular Change Detection in Ukrainian Forest Ecosystem With Sentinel-2

The logging is the leading cause for the reduction in the forest area in the world. At the same time, the number of forest clearcuts continues to grow. However, despite the massive scale, such incidents are difficult to track in time. As a result, huge areas of forests are gradually being cut down. Therefore, there is a need for regular and effective monitoring of changes in forest cover. The multitemporal data sources like Copernicus Sentinel-2 allow enhancing the potential of monitoring the Earth's surface and environmental dynamics including forest plantations. In this article, we present a baseline U-Net model for deforestation detection in the forest-steppe zone. Training and evaluation are conducted on our own dataset created on Sentinel-2 imagery for the Kharkiv region of Ukraine (31 400 km <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ). As a part of the research, we present several models with the ability to work with time-dependent imagery. The main contribution of this article is to provide a baseline model for the forest change detection inside Ukraine and improve it adding the ability to use several sequential images as an input of the segmentation model.

A saliency-guided neighbourhood ratio model for automatic change detection of SAR images

ABSTRACT In this paper, a new approach to the saliency-guided neighbourhood ratio method has been proposed to change-detection of Synthetic Aperture Radar (SAR) images. Two methods, the saliency-guided and the neighbourhood ratio generate a vector as a feature to classify the image difference of SAR images into changed and unchanged classes by using a clustering method such as k-means clustering. Therefore, in this study, the vectors obtained from the saliency-based neighbourhood ratio method was used as separate properties in the k-means clustering algorithm to classify the area into two classes, changed and unchanged region. Furthermore, the saliency-guided method uses a threshold value to generate a saliency map that is selected by trial and error for each data set. In this study, the Otsu threshold histogram method was used to calculate the threshold, which leads to higher accuracy of the method and the automatic selection of the threshold at a higher rate. Moreover, in the neighbourhood ratio model, the probability distribution function image was used instead of the average ratio image. Next, the proposed method has been implemented in three SAR data sets of Bern, Ottawa, and Yellow River China as ground truth data to evaluate the accuracy of the model. The results of the proposed model show that the kappa coefficient (κ) of the Berne data set is increased from 0.8705 in the saliency-guided model and 0.8590 in the neighbourhood ratio method to 0.8821. Also, for the Ottawa data set, the κ of the proposed model has been raised by 0.1 and 0.3190 compared to the saliency-guided and neighbourhood ratio methods, respectively. Finally, the accuracy of the proposed model in the Yellow River data set, respectively, 0.45 and 0.57, is improved than saliency-guided and neighbourhood ratio models.

Partly Uncoupled Siamese Model for Change Detection from Heterogeneous Remote Sensing Imagery

This paper addresses the problematic of detecting changes in bitemporal heterogeneous remote sensing image pairs. In different disciplines, multimodality is the key solution for performance enhancement in a collaborative sensing context. Particularly, in remote sensing imagery there is still a research gap to fill with the multiplication of sensors, along with data sharing capabilities, and multitemporal data availability. This study is aiming to explore the multimodality in a multi-temporal set-up for a better understanding of the collaborative sensor wide information completion; we propose a pairwise learning approach consisting on a pseudo-Siamese network architecture based on two partly uncoupled parallel network streams. Each stream represents itself a Convolutional Neural Network (CNN) that encodes the input patches. The overall Change Detector (CD) model includes a fusion stage that concatenates the two encodings in a single multimodal feature representation which is then reduced to a lower dimension using fully connected layers and finally a loss function based on the binary cross entropy is used as a decision layer. The proposed pseudo-Siamese pairwise learning architecture allows to the CD model to capture the spatial and the temporal dependencies between multimodal input image pairs. The model processes the two multimodal input patches at one-time under different spatial resolutions. The evaluation performances on different real multimodal datasets reflecting a mixture of CD conditions with different spatial resolutions, confirm the effectiveness of the proposed CD architecture.

A generic composite measure of similarity between geospatial variables

The comparison between spatial or temporal patterns is often needed for model evaluation and change detection in ecological studies. The statistics developed for image quality assessment, such as the structural similarity index (SSIM) and the composite similarity measure based on means, standard deviations, and correlation coefficient (CMSC), have been introduced for comparing ecological patterns. However, these measures can be applied only when a positive relationship is expected between patterns having the same scale. We propose a new index, generic composite similarity measure (GCSM), to meet a wide range of potential applications. A set of numerical experiments was performed to illustrate the properties of GCSM in comparison with SSIM and CMSC. Two case studies were conducted examining the (dis)agreement between two products of gross primary production (GPP), and the relative (dis)similarity between GPP and precipitation, respectively. GCSM has advantages over both SSIM and CMSC, including higher sensitivity and the ability to quantify the dissimilarity, which cannot be properly revealed with the latter two indices. The normalization preprocessing constructs universal criteria for assessing the relative (dis)similarity between patterns having unequal scales. The GCSM, overcoming the limitations of preexisting composite measures in quantifying the similarity or dissimilarity between patterns, would aid assessment of heterogeneous relationship between ecological factors over space or time.

Two multivariate online change detection models

ABSTRACT Online change point detection methods monitor changes in the distribution of a data stream. This article discusses two non-parametric online change detection methods based on the energy statistics and Mahalanobis depth. To apply the energy statistic, we use sliding-window algorithm with efficient training and updating procedures. For Mahalanobis depth, we propose an algorithm to train the threshold with desired protective ability against false alarms and discuss factors that have an influence on the threshold. Numerical studies evaluate and compare the performance of the proposed models with three existing methods to detect changes in the mean and variability of a data stream. The methods are applied to detecting changes in the flowing volume of the Mississippi River.

BUILDING CHANGE DETECTION FROM BITEMPORAL AERIAL IMAGES USING DEEP LEARNING

Abstract. Automatic building change detection has become a topical issue owing to its wide range of applications, such as updating building maps. However, accurate building change detection remains challenging, particularly in urban areas. Thus far, there has been limited research on the use of the outdated building map (the building map before the update, referred to herein as the old-map) to increase the accuracy of building change detection. This paper presents a novel deep-learning-based method for building change detection using bitemporal aerial images containing RGB bands, bitemporal digital surface models (DSMs), and an old-map. The aerial images have two types of spatial resolutions, 12.5 cm or 16 cm, and the cell size of the DSMs is 50 cm × 50 cm. The bitemporal aerial images, the height variations calculated using the differences between the bitemporal DSMs, and the old-map were fed into a network architecture to build an automatic building change detection model. The performance of the model was quantitatively and qualitatively evaluated for an urban area that covered approximately 10 km2 and contained over 21,000 buildings. The results indicate that it can detect the building changes with optimum accuracy as compared to other methods that use inputs such as i) bitemporal aerial images only, ii) bitemporal aerial images and bitemporal DSMs, and iii) bitemporal aerial images and an old-map. The proposed method achieved recall rates of 89.3%, 88.8%, and 99.5% for new, demolished, and other buildings, respectively. The results also demonstrate that the old-map is an effective data source for increasing building change detection accuracy.

Selective Adversarial Adaptation-Based Cross-Scene Change Detection Framework in Remote Sensing Images

Supervised change detection methods always face a big challenge that the current scene (target domain) is fully unlabeled. In remote sensing, it is common that we have sufficient labels in another scene (source domain) with a different but related data distribution. In this article, we try to detect changes in the target domain with the help of the prior knowledge learned from multiple source domains. To achieve this goal, we propose a change detection framework based on selective adversarial adaptation. The adaptation between multisource and target domains is fulfilled by two domain discriminators. First, the first domain discriminator regards each scene as an individual domain and is designed for identifying the domain to which each input sample belongs. According to the output of the first domain discriminator, a subset of important samples is selected from multisource domains to train a deep neural network (DNN)-based change detection model. As a result, not only the positive transfer is enhanced but also the negative transfer is alleviated. Second, as for the second domain discriminator, all the selected samples are thought from one domain. Adversarial learning is introduced to align the distributions of the selected source samples and the target ones. Consequently, it further adapts the knowledge of change from the source domain to the target one. At the fine-tuning stage, target samples with reliable labels and the selected source ones are used to jointly fine-tune the change detection model. As the target domain is fully unlabeled, homogeneity- and boundary-based strategies are exploited to make the pseudolabels from a preclassification map reliable. The proposed method is evaluated on three SAR and two optical data sets, and the experimental results have demonstrated its effectiveness and superiority.

Mitigating systematic error in topographic models for geomorphic change detection: accuracy, precision and considerations beyond off‐nadir imagery

AbstractUnmanned aerial vehicles (UAVs) and structure‐from‐motion photogrammetry enable detailed quantification of geomorphic change. However, rigorous precision‐based change detection can be compromised by survey accuracy problems producing systematic topographic error (e.g. ‘doming’), with error magnitudes greatly exceeding precision estimates. Here, we assess survey sensitivity to systematic error, directly correcting topographic data so that error magnitudes align more closely with precision estimates. By simulating conventional grid‐style photogrammetric aerial surveys, we quantify the underlying relationships between survey accuracy, camera model parameters, camera inclination, tie point matching precision and topographic relief, and demonstrate a relative insensitivity to image overlap. We show that a current doming‐mitigation strategy of using a gently inclined (&lt;15°) camera can reduce accuracy by promoting a previously unconsidered correlation between decentring camera lens distortion parameters and the radial terms known to be responsible for systematic topographic error. This issue is particularly relevant for the wide‐angle cameras often integrated into current‐generation, accessible UAV systems, frequently used in geomorphic research. Such systems usually perform on‐board image pre‐processing, including applying generic lens distortion corrections, that subsequently alter parameter interrelationships in photogrammetric processing (e.g. partially correcting radial distortion, which increases the relative importance of decentring distortion in output images). Surveys from two proglacial forefields (Arolla region, Switzerland) showed that results from lower‐relief topography with a 10°‐inclined camera developed vertical systematic doming errors &gt; 0·3 m, representing accuracy issues an order of magnitude greater than precision‐based error estimates. For higher‐relief topography, and for nadir‐imaging surveys of the lower‐relief topography, systematic error was &lt; 0·09 m. Modelling and subtracting the systematic error directly from the topographic data successfully reduced error magnitudes to values consistent with twice the estimated precision. Thus, topographic correction can provide a more robust approach to uncertainty‐based detection of event‐scale geomorphic change than designing surveys with small off‐nadir camera inclinations and, furthermore, can substantially reduce ground control requirements. © 2020 The Authors. Earth Surface Processes and Landforms published by John Wiley &amp; Sons Ltd

A Novel Active Contours Model for Environmental Change Detection from Multitemporal Synthetic Aperture Radar Images

In this paper, we propose a novel approach based on the active contours model for change detection from synthetic aperture radar (SAR) images. In order to increase the accuracy of the proposed approach, a new operator was introduced to generate a difference image from the before and after change images. Then, a new model of active contours was developed for accurately detecting changed regions from the difference image. The proposed model extracts the changed areas as a target feature from the difference image based on training data from changed and unchanged regions. In this research, we used the Otsu histogram thresholding method to produce the training data automatically. In addition, the training data were updated in the process of minimizing the energy function of the model. To evaluate the accuracy of the model, we applied the proposed method to three benchmark SAR data sets. The proposed model obtains 84.65%, 87.07%, and 96.26% of the Kappa coefficient for Yellow River Estuary, Bern, and Ottawa sample data sets, respectively. These results demonstrated the effectiveness of the proposed approach compared to other methods. Another advantage of the proposed model is its high speed in comparison to the conventional methods.

Detecting lane change maneuvers using SHRP2 naturalistic driving data: A comparative study machine learning techniques

Lane change has been recognized as a challenging driving maneuver and a significant component of traffic safety research. Developing a real-time continuous lane change detection system can assist drivers to perform and deal with complex driving tasks or provide assistance when it is needed the most. This study proposed trajectory-level lane change detection models based on features from vehicle kinematics, machine vision, roadway characteristics, and driver demographics under different weather conditions. To develop the models, the SHRP2 Naturalistic Driving Study (NDS) and Roadway Information Database (RID) datasets were utilized. Initially, descriptive statistics were utilized to investigate the lane change behavior, which revealed significant differences among different weather conditions for most of the parameters. Six data fusion categories were introduced for the first time, considering different data availability. In order to select relevant features in each category, Boruta, a wrapper-based algorithm was employed. The lane change detection models were trained, validated, and comparatively evaluated using four Machine Learning algorithms including Random Forest (RF), Support Vector Machine (SVM), Artificial Neural Network (ANN), and eXtrem Gradient Boosting (XGBoost). The results revealed that the highest overall detection accuracy was found to be 95.9 % using the XGBoost model when all the features were included in the model. Moreover, the highest overall detection accuracy of 81.9 % using the RF model was achieved considering only vehicle kinematics-based features, indicating that the proposed model could be utilized when other data are not available. Furthermore, the analysis of the impact of weather conditions on lane change detection suggested that incorporating weather could improve the accuracy of lane change detection. In addition, the analysis of early lane change detection indicated that the proposed algorithm could predict the lane changes within 5 s before the vehicles cross the lane line. The developed detection models could be used to monitor and control driver behavior in a Cooperative Automated Vehicle environment.

Semantic classification and hash code accelerated detection of design changes in BIM models

As the design process of a building or infrastructure is complex with iterative steps and multidisciplinary collaboration, design changes are inevitable and efficient change detection based on Building Information Modeling (BIM) is essential for design collaboration and quality control. Although some detection methods have been proposed to tackle the problem, majority of them regard syntactic data changes as design changes, which always yield incorrect and meaningless results. For instance, exchanging the locations of two identical columns is meaningless from the viewpoint of a designer, which the current methods take it as a design change. Otherwise, additional time is needed to avoid these issues, thus leading to high time consumption. To address this problem, we first propose a semantic classification of design changes with three categories of data changes and three levels of design changes. Then, an efficient detection method considering the proposed semantic classification of design changes is proposed with hash-code-based acceleration. After implementing the proposed and existing methods, the detection performance of our method is compared with the performances of existing methods and invited designers based on 4 groups of BIM models. Results show that our method could: 1) detect semantic changes with 100% correctness while reducing up to 98.1% of the detection time of existing methods, and 2) improve the average detection rate of designers from 81.3% to 100% while saving 99.88% of the design reviewing time (detecting 7 changes from 5800 instances). This research contributes to the body of knowledge a new definition of design changes and an approach to efficient design change detection with hash-code-based acceleration. Meanwhile, the proposed classification of design changes and developed BIM models in this research can also be taken as a baseline for developing and validating new detection methods.

An automatic change detection method for monitoring newly constructed building areas using time-series multi-view high-resolution optical satellite images

Automatically monitoring newly constructed building areas (NCBAs) is essential for efficient land resource management and sustainable urban development, particularly in the rapidly urbanizing country of China. In this regard, time-series multi-view high-resolution optical satellite images can provide fine spatial details for clearly characterizing NCBAs, but this leads to great heterogeneity and complexity, owing to the high spectral variation, complicated imaging conditions, and different viewing angles. Moreover, to date, the vertical features and time-series information from these images have not been fully exploited for urban change detection. In this paper, our primary objective is to automatically detect the presence of NCBAs, and meanwhile, to investigate the feasibility of identifying their change timing using time-series multi-view ZY-3 high-resolution satellite images. To this aim, we propose an automatic change detection method consisting of three components: 1) firstly, we jointly use planar-vertical features to delineate the NCBAs; 2) object-based temporal correction is subsequently applied to improve the spatiotemporal consistency of the features; and 3) finally, a multi-temporal change detection model is used to simultaneously capture the NCBAs and the change timing. We applied the method on two urban fringe areas of Beijing (7 multi-temporal image sets) and Shanghai (7 multi-temporal image sets), respectively, which are cities that have been experiencing rapid urbanization. The experimental results confirmed the effectiveness of the proposed method. For both study areas, the F-score values reached nearly 90% in terms of NCBA detection, and with respect to the change timing, the overall accuracies with a one-year tolerance strategy reached around 92%. The joint use of the planar-vertical features and the inclusion of multi-temporal images make the proposed method a promising approach for automatically providing the spatiotemporal information of NCBAs in practical applications.

Deciphering Circular Anthropogenic Anomalies in PALSAR Data—Using L-Band SAR for Analyzing Archaeological Features on the Steppe

Synthetic aperture radar has been employed for archaeological purposes for nearly forty years: nonetheless, its application among archaeological practitioners has remained limited. We analyzed circular anthropogenic anomalies in a steppe environment in PALSAR-2 data, which appeared as a homogeneous group of signatures. Each anomaly was examined using additional SAR and optical data, as well as investigated through extensive ground truth and, in one case, excavation. We found the anomalies to originate from a wide range of processes and structural characteristics showing the non-intuitive complexity of SAR data interpretation. We found that this is likely the reason for the limited application SAR has seen within the archaeological community. In order to improve the usage of SAR for archaeological purposes beyond change detection and digital elevation models, specific products that are more readily understandable and superior to optical data in a narrow frame of application should be developed.