Deep learning-based hyperspectral image (HSI) classification models typically utilize multiple feature extraction layers to learn the features of land covers. Nevertheless, they encounter challenges, e.g., 1) Transformers require substantial computational resources, and 2) these layers are carefully assembled and designed. Recently, large kernel convolutional neural networks (LKCNNs) show excellent performance in natural visual tasks. To tackle these limitations and explore the capability of LKCNNs for HSI classification, we present a novel simple and powerful multi-scale large kernel convolutional neural network architecture (MSLKCNN) with the largest kernel size as large as 15 × 15, in contrast to commonly used 3 × 3, for HSI classification. MSLKCNN avoids these specialized designs, comprising a noise suppression module (NSM) and a multi-scale large kernel convolution (MSLKC). Specifically, NSM is first used to suppress the noise and reduce the number of the bands before extracting the features. Then, MSLKC, as the only feature extraction layer of MSLKCNN, joints three parallel convolutions to capture the features of various types (i.e. spectral, spectral-spatial) and ranges (i.e., small local, larger local, and global) from the dimension of scale: (C1) convolution with a kernel size of 1 × 1 is used to extract spectral features; (C2) multi-scale large kernel depthwise separable convolution (MLKDC) is proposed to learn the spectral-spatial features of different ranges including short-range, middle-range, and long-range; and (C3) multi-scale dilated depthwise separable convolution (MDDC) is designed to aggregate the spectral-spatial features between land covers at various distances. Extensive experimental results on three public HSI datasets demonstrate the competitiveness of the proposed MSLKCNN compared with several state-of-the-art methods.

Current multi-polarization Time-Series Interferometric Synthetic Aperture Radar (TS-InSAR) approaches conduct polarimetric optimization during the coherent scatterer selection and interferogram generation stages. However, these techniques often encounter limitations when dealing with certain types of targets, particularly when the target scattering phase center positions differ across polarizations. Consequently, the temporal coherence of the optimized interferometric phase can sometimes be worse than that of the classical single-polarization technique. This work aims to enhance these techniques by fine-tuning the TS-InSAR workflow and selecting the most suitable observation from various polarization channels based on temporal coherence. Experimental results show that the proposed method increases the number of persistent scatterers and offers more reliable deformation monitoring than both single-polarization and existing multi-polarized methods.

Pre-disaster information storage is crucial for effective disaster response. The discussion regarding deep learning-based Light Detection and Ranging (Lidar) semantic segmentation technology for indoor small items has been ongoing in recent years. However, the methods applicable to large-scale outdoor Lidar datasets for pre-disaster information storage remain limited. This study aims to propose a novel deep learning-based network for city-scale Lidar semantic segmentation to support pre-disaster information storage, called channel attention and normal-based local feature aggregation network (CNLNet). This network is designed to segment common urban land cover objects, including buildings and vegetation. This network incorporates surface normal information and the channel attention mechanism into the RandLA-Net backbone. Ablation studies have been devised to assess the performance of these two features. During the pre-processing step, color information from optical images is fused with Lidar data. The findings demonstrate that CNLNet can enhance the accuracy of the RandLA-Net backbone by improving mIoU at least 1-2%. Including one of these two features also contributes to the backbone’s improved accuracy. Notably, CNLNet outperforms other well-known networks in terms of accuracy with the test of the public Sementic3D dataset. The study further reveals that the proposed network excels in building segmentation, a crucial facet of pre-disaster information storage. Moreover, the results show that spatial resolution, whether at 0.5m or 10m per pixel for optical images, has limited influence on outcomes. One theoretical contribution of this study is the demonstration of the advantages of integrating either surface normal information or a channel attention mechanism to enhance large-scale outdoor Lidar semantic segmentation. Labeled Lidar datasets have been created for training. The practical contribution is that it can optimize disaster response by efficiently facilitating pre-disaster information storage.

The implementation of a wide-land deformation monitoring program in Eastern states of Australia, characterized by an extensive landmass and extensive coastlines, necessitates the generation of deformation maps with a significantly larger spatial extent. A pragmatic approach for achieving such coverage involves the integration of multiple Interferometric Synthetic Aperture Radar (InSAR) time series results, encompassing various tracks and frames. The Sentinel-1 satellite constellation is facilitating the generation of expansive ground surface deformation, while the initial challenge in obtaining such product arises from the varying spatial coverage of the Sentinel-1 data. To overcome this obstacle, we proposed a novel Sentinel-1 image definition to facilitate consistent interferometric processing. Nevertheless, prior to the formation of any large displacement map, the inconsistency between different SAR image scenes must be addressed. Several factors contribute to these inconsistencies in observations, such as differences in the angle of observation at the overlapping regions of adjacent image tracks, and imprecise estimations within the dataset itself (e.g., burst or swath discontinuities). This study delves into solutions for these challenges, specifically: 1) introducing a computer-vision-based algorithm for InSAR dataset quality assessment, and 2) proposing a global least square mosaicking procedure for the amalgamation of tiles from multi-tracks and frames. Statistical analyses show that better accuracy can be achieved. The mosaicking InSAR product demonstrates the capacity to quantify ground surface changes, which also exhibits a correlation with other geological layers, and the subsidence range of [-7, 10] mm/yr shows a strong association with clay content levels ranging from 10% to 30% along the Darling River.

The Sentinel-1 satellite constellation provides temporally dense and high spatial resolution Synthetic Aperture Radar (SAR) imagery. The open data policy and global coverage of Sentinel-1 enables applications that require long time series, such as modelling surface deformation rates from Interferometric Synthetic Aperture Radar (InSAR) data, to be scaled-up and generate continent-scale products. In this paper, we present a workflow to generate interferometric products from the Copernicus Australasia Regional Data Hub archive of Single Look Complex (SLC) Sentinel-1 data, with the objective to derive a continent-wide ground surface deformation map for Australia. The proposed workflow can be used to generate Sentinel-1 InSAR Analysis Ready Data (ARD) suitable for ongoing monitoring of ground deformation.

Early detection of bushfire plays a crucial role in firefighting, fire modelling, and minimising losses of human lives and properties. However, current bushfire monitoring systems have an intrinsic shortcoming because only temperature difference between neighboring pixels is exploited. This paper proposes to also examine a range of other changes occur when a bushfire is ignited, for example, a reduction of vegetation cover, volume scattering of bush and trees, as well as height of vegetation. All of these can be readily measured by optical and radar satellites already in orbits in near real-time, that is, less than two hours after a satellite overpass. Cross-correlation of these measurements has the potential to significantly reduce false alarm of a bushfire, while improving the early detection and measurement of fire spots, and hence make the system much more robust. A case study near Sydney is included here based on Sentinel-1 SAR and Sentinel-2 optical satellite data collected on 10 and 11 October 2020, respectively. This research is a major step forward towards the operational and synergetic use of optical and SAR satellites in bushfire monitoring.

In this paper, the 2017 Maoxian landslide in Sichuan province, China was investigated by using multi-temporal SAR and optical datasets. The pre- and post-failure scars of the landslide were depicted by using the K-means classification of the Normalized Difference Vegetation Index (NDVI) maps. Two stacks of ascending and descending Sentinel-1A/B C-band SAR images were applied to explore the pre-failure characteristics of this event. Advanced time-series InSAR analysis was processed to detect the pre-failure movements of this event, with identifying the spatial-temporal pattern and evolution of the source area of the Maoxian landslide. The maximum line-of-sight (LOS) deformation rate measured over the slope surface was up to -30 mm/year in the descending track, with only -18 mm/year for the ascending track. Most importantly, an obvious acceleration was detected from the time series analysis of selected measurement points at the source area before the occurrence of failure. By comparing the TS-InSAR result with the precipitation record over this region, it is evidenced that heavy rainfall might be the major triggering factor of the Maoxian Landslide.

In this study, the long term land subsidence (~ 15 years) in the Mexico City, Mexico mapped using two in-house InSAR methods GEOS (Geoscience and Earth Observing Systems Group)-ATSA (Advance Time-Series Analysis) and GEOS-SBAS (Small Baseline Subset) has been presented. An IDW (Inverse Distance Weighted)-based integration module and MLR (maximum likelihood regression)-based M-estimator are introduced to further enhance these two methods. The land subsidence was continuously mapped using Envisat (2004 - 2007), ALOS-1 (2007 - 2011), CSK (2011 - 2014), ALOS-2 (2014 - 2018), and Sentinel-1 (2015 - 2017) datasets. A comparison between InSAR time series and GPS measurement shows that the subsidence rates were consistent over 2004 - 2018, and five evidences were given to support this argument. The 15-year accumulated subsidence map was generated and the maximum subsidence over 4.5 m was found. By comparing the InSAR result with land use map, it has been found that there are some relations between the local subsidence rate and land use type while residential usage and consumption of the groundwater has quite significant contribution to the local subsidence rate.

This paper addresses the forest height inversion based on the modified RMoG model using repeat-pass Pol-InSAR data. The linear variance of the Gaussian motion distribution in the RMoG model is replaced by a linear standard deviation to describe the volumetric motion heterogeneity and related coherence function is deduced based on that. Furthermore, a forest parameter inversion algorithm is proposed based on this modified model and its performance is investigated with L-band repeat-pass ALOS-1 quad-polarization data acquired over German forest site with temporal baselines of 46 days. Inversion results indicates that in comparison with the traditional RVoG and RMoG methods, the modified method can reduce 27.73% and 8.57% of the overestimation errors caused by the temporal decorrelation.

Continuous research has been conducted in Bandung City, West Java province, Indonesia over the past two decades. Previous studies carried out in a regional-scale might be useful for estimating the correlation between land subsidence and groundwater extraction, but inadequate for local safety management as subsidence may vary over different areas with detailed characters. This study is focused primarily on subsidence phenomenon in local, and patchy scales, respectively, with Sentinel-1 and ALOS-2 dataset acquired from September 2014 to July 2017. In order to understand the subsidence in a more systematic way, six 10-cm subsidence zones have been selected known as Zone A to F. Further analyses conducted over multiple scales show that industrial usage of groundwater is not always the dominant factor that causes the land subsidence and indeed it does not always create large land subsidence either. Regions experiencing subsidence is due to a combined impact of a number of factors, e.g., residential, industrial or agricultural activities.

SqueeSAR, also known as advanced time-series interferometric synthetic aperture radar (ATS-InSAR) method, is a significant improvement of conventional persistent scatterer InSAR (PSInSAR), whereby the concepts of distributed scatterer (DS) and persistent scatterer (PS) are first been introduced, respectively. It is worth noting that during the measurement pixel selection, it is inevitable that a number of PS can be categorized as DS as well, hence resulting in common PS-DS pixels. In order to understand the consistency among these common PS-DS pixels with PSInSAR and ATS-InSAR methods, statistical analyses are conducted with 10 real InSAR image stacks in this letter. The relationship between the goodness-of-fit value and four main factors, including root-mean-square difference, DS percentage, PS-DS/PS ratio, and PS-DS/DS ratio, is studied. It is concluded that Sentinel-1-based TS-InSAR can be less influenced by the goodness-of-fit threshold in comparison with the counterpart result of ALOS-1 under the same parameter setting; finally, conclusions for the threshold settings are given.

Advanced Time series InSAR (ATS-InSAR) is generally refer to those TS-InSAR methods with an external Distributed Scatterer (DS) selection module, e.g. SqueeSAR and GEOS-ATSA. It is being known as a very efficient tool for monitoring the ground deformation over suburban or even non-urban regions with great success. However, within Appin Colliery, which is located in the southeastern corner of the Southern Coalfield, New South Wales (NSW), Australia. C-band ASAR based ATS-InSAR failed to produce reasonable outcome due to the underground mining effect. This paper presents a modified ATS-InSAR method for mapping the ground deformation over underground mining region. Firstly, traditional reliable DS pixels and Persistent Scatterer (PS) pixels are selected to form the initial triangular irregular network (TIN) reference network. Then the ground deformation and DEM error with respect to these Measurement Scatterer (MS) pixels are solved through a robust regression estimator. Due to the losses of coherence, the general underground mining pattern cannot be formed when using C-band image stacks. Therefore, in order to achieve the best detail, modified MS pixel selection method is conducted by including less reliable MS pixels based on a weighted least square method. Moreover, final result proved to be efficient to offer sufficient information to associated councils and department for risk management purpose.

Beijing metropolitan, the capital city of China, has suffered from the groundwater-induced subsidence since the late 1950s. The previous researches with respect to the ground deformation in Beijing City mainly focused on the time period before the year of 2014. To study the recent evolution of ground deformation, twenty-four C-band Sentinel-1A/B images (June 2015-November 2016) along with nineteen L-band ALOS-1 PALSAR images (June 2007-January 2011) are analysed in this research. As the optical-based classification result indicates that approximate 42% of the processed area is covered by rural land-use type, of which the main uses are farmland and grasslands, therefore, to achieve the best detail over both urban/non-urban regions, a Distributed Scatterer based TS-InSAR is implemented to provide the timely information for ground deformation assessment. It is worth noting that even though both descending and ascending time series InSAR (TS-InSAR) deformation products are available, the vertical deformation is not estimated due to that the temporal period between these two products are not the same. Moreover, the general subsidence patterns from -20 mm·yr-1to -120 mm·yr-1section over these two datasets are correlated to some extent, which suggests that the subsiding trend is still continuing. However, there is a vast difference between - 120 mm·yr-1to -150 mm·yr-1section, which reveals the fact that the maximum subsidence rates over some regions are decelerating.

This paper reports the findings based on ALOS-1 and GRACE satellite data for the purpose of monitoring land surface subsidence due to groundwater extraction in the Ordos Basin, China. Twenty ALOS-1 PALSAR data acquired between 8 January 2007 and 19 January 2011 are utilized in the time-series InSAR interferometry (TS-InSAR) analysis while the total water storage observations derived from the Gravity Recovery and Climate Experiment (GRACE) satellite data are integrated with hydrological modeling results (for soil moisture modelling) to estimate the groundwater depletion rate. The outcome shows that the total mean subsidence measured from TS-InSAR is about -6.8 mm yr-1in vertical direction while the groundwater depletion rate is about -4.2 mm yr-1between 17 December 2006 and 15 December 2010. Since in general every 1 m drop in groundwater level could lead to land subsidence of about 5 to 50 mm, and the total subsidence in Ordos Basin is mainly due to groundwater extraction and underground mining activities. The experiment result shows that the total land subsidence is mainly induced by underground mining.

This paper reports the findings on monitoring land subsidence in the south of Sydney, Australia, using data from ALOS-1 and ENVISAT satellites. Twenty-three L-band ALOS-1 PALSAR scenes acquired between 29 June 2007 and 07 January 2011 and twenty-six C-band ENVISAT ASAR image acquired between 09 July 2007 and 06 September 2010 are analysed in this research. Since the city of Wollongong and the town of Appin underground mining site are two most interested regions in the south of Sydney for this study, a new strategy is proposed to select measurement points according to different geophysical information in order to achieve the best detailed deformation mapping. In this paper, both descending and ascending time series InSAR (TS-InSAR) deformation products are utilised to estimate the vertical deformation in Wollongong city area. The highest vertical deformation less than -0.8 cm yr-1is detected. It is worth noting that TS-InSAR analysis to C-band satellite dataset cannot generate reasonable result in Appin underground mining site due to its limited dynamic range of detectable subsidence.

Land displacement in Ordos, China, between 8 January 2007 and 19 January 2011 was mapped using L-band ALOS PALSAR data. Twenty ALOS PALSAR scenes acquired were utilised to generate both PSI and DInSAR results. Several locations in the eastern Ordos experiencing rapid land subsidence were identified including Huo Luo Wan coalmine and Qu Jia Liang coalmine. The subsidence rates ranging from -35mm/year to 35 mm/year were detected. The comparison between PSI and DInSAR outcomes, although showing good agreement in general, reveals some gaps in PSI map near Qu Jia Liang coalmine mainly due to rapid changes within the four-year period. Six successive DInSAR results were exploited to generate time series deformation map, with selected points being analysed and the reason being given for the formation of gaps as well. The DInSAR deformation measurements were then converted into time series velocity maps and integrated with PSI outcome to generate a final product.