High Resolution Optical Imagery Research Articles

Assessment of Spatio-Temporal Landscape Changes from VHR Images in Three Different Permafrost Areas in the Western Russian Arctic

Our study highlights the usefulness of very high resolution (VHR) images to detect various types of disturbances over permafrost areas using three example regions in different permafrost zones. The study focuses on detecting subtle changes in land cover classes, thermokarst water bodies, river dynamics, retrogressive thaw slumps (RTS) and infrastructure in the Yamal Peninsula, Urengoy and Pechora regions. Very high-resolution optical imagery (sub-meter) derived from WorldView, QuickBird and GeoEye in conjunction with declassified Corona images were involved in the analyses. The comparison of very high-resolution images acquired in 2003/2004 and 2016/2017 indicates a pronounced increase in the extent of tundra and a slight increase of land covered by water. The number of water bodies increased in all three regions, especially in discontinuous permafrost, where 14.86% of new lakes and ponds were initiated between 2003 and 2017. The analysis of the evolution of two river channels in Yamal and Urengoy indicates the dominance of erosion during the last two decades. An increase of both rivers’ lengths and a significant widening of the river channels were also observed. The number and total surface of RTS in the Yamal Peninsula strongly increased between 2004 and 2016. A mean annual headwall retreat rate of 1.86 m/year was calculated. Extensive networks of infrastructure occurred in the Yamal Peninsula in the last two decades, stimulating the initiation of new thermokarst features. The significant warming and seasonal variations of the hydrologic cycle, in particular, increased snow water equivalent acted in favor of deepening of the active layer; thus, an increasing number of thermokarst lake formations.

A Hidden Markov Tree Model for Flood Extent Mapping in Heavily Vegetated Areas based on High Resolution Aerial Imagery and DEM: A Case Study on Hurricane Matthew Floods

ABSTRACT Flood extent mapping plays a crucial role in disaster management and national water forecasting. In recent years, high-resolution optical imagery becomes increasingly available with the deployment of numerous small satellites and drones. However, analysing such imagery data to extract flood extent poses unique challenges due to the rich noise and shadows, obstacles (e.g. tree canopies, clouds), and spectral confusion between pixel classes (flood, dry) due to spatial heterogeneity. Existing machine-learning techniques often focus on spectral and spatial features from raster images without fully incorporating the geographic terrain within classification models. In contrast, we recently proposed a novel machine-learning model called geographical hidden Markov tree (HMT) that integrates spectral features of pixels and topographic constraint from Digital Elevation Model (DEM) data (i.e. water flow directions) in a holistic manner. This paper evaluates the model through case studies on high-resolution aerial imagery from National Oceanic and Atmospheric Administration (NOAA) National Geodetic Survey (NGS) together with DEM. Three scenes are selected in heavily vegetated floodplains near the cities of Grimesland and Kinston in North Carolina during Hurricane Matthew floods in 2016. Results show that the proposed HMT model outperforms several state of the art machine-learning algorithms (e.g. random forests, gradient-boosted model) by an improvement of F-score (the harmonic mean of the user’s accuracy and producer’s accuracy) from around 70% to 80% to over 95% on our datasets.

The impact of wildfires on the Lake Skadar/Shkodra environment

Every year wildfires bring damages to the environment of Montenegro, and, in particular, to the Lake Skadar/Shkodra environment. Wildfires are caused by dry and hot weather which is established yearly from July to September, and is likely linked to the regional climate change. In this study we used high resolution optical imagery (OLI Landsat-8 and MSI Sentinel-2A, -2B) to analyze concentrations of suspended particulate matter and chlorophyll-a in Lakes Skadar/Shkodra and Šas in order to understand the potential impact of wildfires on the lake ecosystem. The surface of burnt areas was calculated for different regions around Lakes Skadar/Shkodra and Šas. Our data showed that SPM values in lake water increase during wildfire. The highest SPM value was observed in the area around the mouth of the left branch of River Morača where the largest areas affected by fires were observed. The SPM concentration in water of Lakes Skadar/Shkodra and Šas generally decreased after rainstorms. The need for implementation of a comprehensive fire management program that includes monitoring of impact of fires on aquatic ecosystems and their biota is emphasized.

Spatial Variability Assessment on The High-resolution Chlorophyll-a Extraction from Landsat 8and Sentinel 2 Imageries in Johor Waters

High resolution Chlorophyll-a (Chl-a) can indicate the trophic status of the water and provide useful information on optical features of water body in water quality monitoring. Remote sensing has great potential to offer the spatial and temporal coverage needed. Over the last decades the Sea WIFS and MODIS were applied, but not suitable due to the low spatial resolution for monitoring Chl-a in coastal area. However, the retrieval of Chl-a in the coastal region is usually challenging due to the other in-water substances regardless of Chl-a, hence resulting in the satellite retrieved Chl-a overestimation. By the advancement of the Sentinel-2 and Landsat 8 satellites, continuous high resolution optical imageries have served for remarkable coastal mapping capability thanks to the spectroscopic capability certain spectral bands and as high as 10-meter spatial resolution. This paper aims to evaluate the performance of the SEADASS and SNAP processor for Chl-a estimation from the Operational Land Imager (OLI)and MultiSpectral Instrument(MSI) data in Johor waters. The representative models, in standard algorithm OC3and C2RCC, were adapted to retrieve Chl-a concentration. The statistical regression has shown that these algorithms give an acceptable Chl-a estimation at medium and high resolution with R2=0.44 from OC3and R2=0.55from C2RCC comparing to the in-situ data. Despite of the spatial, temporal and spectral variability, this paper shows that OLI and MSI could provide Chl-a mapping capability at suitable Chl-a estimation techniques.

Land motion detection in central Rwanda using small baseline subset interferometry

Synthetic Interferometric Aperture Radar (InSAR) is a cost-effective technique for observing the Earth's surface, topography and ground motion deformations of the Earth. However, little work has been done in developing countries to apply this technique to the monitoring of natural disasters that strongly affect their socioeconomic condition. In this work, we apply the Small BAseline Subset (SBAS) technique in central Rwanda. This article presents the monitoring of earthflows in central Rwanda using 26 Sentinel–1 radar data extended from 2014 to 2018. The site investigation by high-resolution optical imagery from Google Maps is compared to the SBAS velocity. The results revealed an active earthflow near a mining area with a mean velocity around -30 to -20 mm/year and an earthflow of -40 mm/year mean velocity in Kigali. The other movements detected were due to anthropogenic and coastal subsidence.

Significant contribution of small icebergs to the freshwater budget in Greenland fjords

Icebergs represent nearly half of the mass loss from the Greenland Ice Sheet and provide a distributed source of freshwater along fjords which can alter fjord circulation, nutrient levels, and ultimately the Meridional Overturning Circulation. Here we present analyses of high resolution optical satellite imagery using convolutional neural networks to accurately delineate iceberg edges in two East Greenland fjords. We find that a significant portion of icebergs in fjords are comprised of small icebergs that were not detected in previously-available coarser resolution satellite images. We show that the preponderance of small icebergs results in high freshwater delivery, as well as a short life span of icebergs in fjords. We conclude that an inability to identify small icebergs leads to inaccurate frequency-size distribution of icebergs in Greenland fjords, an underestimation of iceberg area (specifically for small icebergs), and an overestimation of iceberg life span.

Testing Urban Flood Mapping Approaches from Satellite and In-Situ Data Collected during 2017 and 2019 Events in Eastern Canada

The increasing frequency of flooding worldwide has driven research to improve near real-time flood mapping from remote-sensing data. Improved automation and processing speed to map both open water and vegetated area flooding have resulted from these research efforts. Despite these achievements, flood mapping in urban areas where a significant number of overall impacts are felt remains a challenge. Near real-time data availability, shadowing caused by manmade infrastructure, spatial resolution, and cloud cover inhibiting optical transmission, are all factors that complicate detailed urban flood mapping needed to inform response efforts. This paper uses numerous data sources collected during two major flood events that impacted the same region of Eastern Canada in 2017 and 2019 to test different urban flood mapping approaches presented as case studies in three separate urban boroughs. Cloud-free high-resolution 3 m PlanetLab optical data acquired near peak-flood in 2019 were used to generate a maximum flood extent product for that year. Approaches using new Lidar Digital Elevation Models (DEM)s and water height estimated from nineteen RADARSAT-2 flood maps, point-based flood perimeter observations from citizen geographic information, and simulated traffic camera or other urban sensor network data were tested and verified using independent data. Coherent change detection (CCD) using multi-temporal Interferometric Wide (IW) Sentinel-1 data was also tested. Results indicate that while clear-sky high-resolution optical imagery represents the current gold standard, its availability is not guaranteed due to timely coverage and cloud cover. Water height estimated from 8 to 12.5 m resolution RADARSAT-2 flood perimeters were not sufficiently accurate to flood adjacent urban areas using a Lidar DEM in near real-time, but all nineteen scenes combined captured boroughs that flooded at least once in both flood years. CCD identified flooded boroughs and roughly captured their flood extents, but lacked timeliness and sufficient detail to inform street-level decision-making in near real-time. Point-based flood perimeter observation, whether from in-situ sensors or high-resolution optical satellites combined with Lidar DEMs, can generate accurate full flood extents under certain conditions. Observed point-based flood perimeters on manmade features with low topographic variation produced the most accurate flood extents due to reliable water height estimation from these points.

ERS-1/2 and Sentinel-1 SAR Data Mining for Flood Hazard and Risk Assessment in Lima, Peru

The coastline environment and urban areas of Peru overlooking the Pacific Ocean are among the most affected by El Niño-Southern Oscillation (ENSO) events, and its cascading hazards such as floods, landslides and avalanches. In this work, the complete archives of the European Space Agency (ESA)’s European Remote-Sensing (ERS-1/2) missions and European Commission’s Copernicus Sentinel-1 constellation were screened to select synthetic aperture radar (SAR) images covering the most severe and recent ENSO-related flooding events that affected Lima, the capital and largest city of Peru, in 1997–1998 and 2017–2018. Based on SAR backscatter color composites and ratio maps retrieved from a series of pre-, cross- and post-event SAR pairs, flooded areas were delineated within the Rímac River watershed. These are mostly concentrated along the riverbanks and plain, where low-lying topography and gentle slopes (≤5°), together with the presence of alluvial deposits, also indicate greater susceptibility to flooding. A total of 409 areas (58.50 km2) revealing change were mapped, including 197 changes (32.10 km2) due to flooding-related backscatter variations (flooded areas, increased water flow in the riverbed, and riverbank collapses and damage), and 212 (26.40 km2) due to other processes (e.g., new urban developments, construction of river embankments, other engineering works, vegetation changes). Urban and landscape changes potentially contributing, either detrimentally or beneficially, to flooding susceptibility were identified and considered in the overall assessment of risk. The extent of built-up areas within the basin was mapped by combining information from the 2011 Global Urban Footprint (GUF) produced by the German Aerospace Center (DLR), the Open Street Map (OSM) accessed from the Quantum GIS (QGIS) service, and 2011–2019 very high-resolution optical imagery from Google Earth. The resulting flooding risk map highlights the sectors of potential concern along the Rímac River, should flooding events of equal severity as those captured by SAR images occur in the future.

Automatic silage maize detection based on phenological rules using Sentinel-2 time-series dataset

ABSTRACT Recently, availability of valuable high spatial and temporal resolution optical satellite imagery such as Sentinel-2 has provided a great opportunity for a wide variety of crop-related studies including crop phenology detection, crop biophysical and biochemical parameter estimation, and crop-type mapping. Most current crop mapping methods focus on supervised classification, demanding a large number of field sampling points to feed the classifier. In this paper, we propose an automatic method for silage maize mapping based on Sentinel-2 time-series data in three agricultural areas in Iran (namely Marvdasht, Abyek, and Mashhad) and an individual area in USA (Tulare County, California). We considered Marvdasht as the training site and other three sites as the test sites. After the preprocessing step, temporal profiles of the Normalized Difference Vegetation Index (NDVI) were computed from Sentinel-2 time-series images in 2017 and 2018. Then, the Savitzky–Golay filter was applied to the NDVI profile for noise reduction. Linear interpolation was also used for the construction of continuous NDVI time-series during the maize growing season. Then, phenological parameters were extracted in Marvdasht, for maize and other crops. After comprehensive analyses, the ratio of the slope from the peak of the greenness to the harvest (SPGH) to the length of the growing season (LOS) was suggested as an appropriate variable for automatic maize detection. Evaluation of this variable in the study sites confirmed its feasibility for automatic maize mapping, with the kappa coefficient values of 0.89, 0.8, 0.9, and 0.8 in Abyek, Marvdasht, Mashhad, and Tulare in 2017, respectively. The performance of the suggested method for the second year (2018) is close to those of 2017 which means that maize detection algorithm results are consistent and robust across years. Moreover, to ascertain high potential of the proposed variable, the results were compared with those of the maximum likelihood (ML) and support vector machine (SVM) classifiers in the test sites. The performance of SVM (with a large number of training samples) was analogous to that of the suggested automatic method in terms of accuracy; however, inferior performance of ML was evident. The results of this research demonstrated the great potential of our proposed phenological method in automatic silage maize mapping.

Integration of high-resolution optical and SAR satellite remote sensing datasets for aboveground biomass estimation in subtropical pine forest, Pakistan.

In this study, we investigate stand-alone and combined Pleiades high-resolution passive optical and ALOS PALSAR active Synthetic Aperture Radar (SAR) satellite imagery for aboveground biomass (AGB) estimation in subtropical mountainous Chir Pine (Pinus roxburghii) forest in Murree Forest Division, Punjab, Pakistan. Spectral vegetation indices (NDVI, SAVI, etc.) and sigma nought HV-polarization backscatter dB values are derived from processing optical and SAR datasets, respectively, and modeled against field-measured AGB values through various regression models (linear, nonlinear, multi-linear). For combination of multiple spectral indices, NDVI, TNDVI, and MSAVI2 performed the best with model R2/RMSE values of 0.86/47.3 tons/ha. AGB modeling with SAR sigma nought dB values gives low model R2 value of 0.39. The multi-linear combination of SAR sigma nought dB values with spectral indices exhibits more variability as compared with the combined spectral indices model. The Leave-One-Out-Cross-Validation (LOOCV) results follow closely the behavior of the model statistics. SAR data reaches AGB saturation at around 120-140 tons/ha, with the region of high sensitivity around 50-130 tons/ha; the SAR-derived AGB results show clear underestimation at higher AGB values. The models involving only spectral indices underestimate AGB at low values (< 60 tons/ha). This study presents biomass estimation maps of the Chir Pine forest in the study area and also the suitability of optical and SAR satellite imagery for estimating various biomass ranges. The results of this work can be utilized towards environmental monitoring and policy-level applications, including forest ecosystem management, environmental impact assessment, and performance-based REDD+ payment distribution.

INCORPORATING INTERFEROMETRIC COHERENCE INTO LULC CLASSIFICATION OF AIRBORNE POLSAR-IMAGES USING FULLY CONVOLUTIONAL NETWORKS

Abstract. Inspired by the application of state-of-the-art Fully Convolutional Networks (FCNs) for the semantic segmentation of high-resolution optical imagery, recent works transfer this methodology successfully to pixel-wise land use and land cover (LULC) classification of PolSAR data. So far, mainly single PolSAR images are included in the FCN-based classification processes. To further increase classification accuracy, this paper presents an approach for integrating interferometric coherence derived from co-registered image pairs into a FCN-based classification framework. A network based on an encoder-decoder structure with two separated encoder branches is presented for this task. It extracts features from polarimetric backscattering intensities on the one hand and interferometric coherence on the other hand. Based on a joint representation of the complementary features pixel-wise classification is performed. To overcome the scarcity of labelled SAR data for training and testing, annotations are generated automatically by fusing available LULC products. Experimental evaluation is performed on high-resolution airborne SAR data, captured over the German Wadden Sea. The results demonstrate that the proposed model produces smooth and accurate classification maps. A comparison with a single-branch FCN model indicates that the appropriate integration of interferometric coherence enables the improvement of classification performance.

Building Extraction in Multitemporal High-Resolution Remote Sensing Imagery Using a Multifeature LSTM Network

Inspired by recent promising developments by deep learning networks, this letter presents a novel approach based on a multifeature long short-term memory (LSTM) network with an optimal unit number for extracting buildings from multitemporal high-resolution optical satellite imagery. The algorithm first extracts specified multifeature data of buildings, including spectral, shape, texture, and indices from multitemporal high-resolution imagery. Next, a designed LSTM network architecture with an optimal unit number is trained using the multifeature data to extract buildings at the pixel level. The proposed network is compared with popular deep learning-based networks, including VGG, U-Net, and ResNet. Finally, postprocessing (e.g., morphological algorithm) is employed to optimize the classification results produced by deep learning. The approach is validated on a newly created data set using Gaofen-2 (GF-2) imagery that contains various buildings on a campus with different materials, shapes, sizes, heights, and directions. Experimental results indicate that the proposed approach outperforms current deep learning-based methods for building extraction and has a potential for practical applications.

Detailed Lacustrine Calving Iceberg Inventory from Very High Resolution Optical Imagery and Object-Based Image Analysis

In the field of iceberg and glacier calving studies, it is important to collect comprehensive datasets of populations of icebergs. Particularly, calving of lake-terminating glaciers has been understudied. The aim of this work is to present an object-based method of iceberg detection and to create an inventory of icebergs located in a proglacial lagoon of San Quintín glacier, Northern Patagonia Icefield, Chile. This dataset is created using high-resolution WorldView-2 imagery and a derived DEM. We use it to briefly discuss the iceberg size distribution and area–volume scaling. Segmentation of the multispectral imagery produced a map of objects, which were classified with use of Random Forest supervised classification algorithm. An intermediate classification product was corrected with a ruleset exploiting contextual information and a watershed algorithm that was used to divide multiple touching icebergs into separate objects. Common theoretical heavy-tail statistical distributions were tested as descriptors of iceberg area and volume distributions. Power law models were proposed for the area–volume relationship. The proposed method performed well over the open lake detecting correctly icebergs in all size bands except 5–15 m2 where many icebergs were missed. A section of the lagoon with ice melange was not reliably mapped due to uniformity of the area in the imagery and DEM. The precision of the DEM limited the scaling effort to icebergs taller than 1.7 m and larger than 99 m2, despite the inventory containing icebergs as small as 4 m2. The work demonstrates viability of object-based analysis for lacustrine iceberg detection and shows that the statistical properties of iceberg population at San Quintín glacier match those of populations produced by tidewater glaciers.

Filtering DSM extraction from Worldview-3 images to DTM using open source software

Today, the Indonesia one map policy has encouraged for producing maps up to scale 1/5.000 – 1/10.000. Therefore, all efforts to accelerate large-scale map production are still needed. One of the challenges is for producing DEM (Digital Elevation Model). DEM data extraction from high-resolution optical satellite imagery is an efficient choice for the production of large-area DEM. This study uses Worldview 3 stereo ready images for DSM (Digital Surface Model) Extraction. Although the DSM data have a high resolution 30cm x 30cm, unfortunately, the DEM extraction techniques from optical images produce DSM (Digital Surface Model) data instead of DTM. While topographic maps generally use DTM data instead of DSM. So, the goal of this study is to use an appropriate procedures method in Open-Source (OS) software towards the generation of Bare-Earth data by filtering DSM. The OS software is SAGA-GIS that it is can be used for any point cloud data. The quality of DTM generated by SAGA-GIS is evaluated by visual interpretation from Worldview 3 orthoimage itself. A good filtering result will remove the non-terrain point. Some independent check point used to evaluate the error in elevation. The error in elevation still less than 2 pixels. Final product is a high resolution DTM which has 30cm × 30cm grid spacing.

Unmanned Aerial Vehicle (UAV)-Based Thermal Infra-Red (TIR) and Optical Imagery Reveals Multi-Spatial Scale Controls of Cold-Water Areas Over a Groundwater-Dominated Riverscape

The forecast of warmer weather, and reduced precipitation and streamflow under climate change makes freshwater biota particularly vulnerable to being exposed to temperature extremes. Given the importance of temperature to regulate vital physiological processes, the availability of discrete cold-water patches (CWPs) in rivers to act as potential thermal refugia is critical to support freshwater ecosystem function. Being able to predict their spatial distribution at riverscape scales is the first step to understanding the capacity to maintain thermal refuges and to inform future river management strategies. Novel Unmanned Aerial Vehicle (UAV)-based Thermal Infra-Red (TIR) imagery technologies provide an opportunity to assess riverscape stream temperature. On the example of a 50 km linear length of the groundwater-dominated Upper Ovens River (Australia), this study presents a methodology addressing critical challenges in UAV-based TIR and optical data acquisition, processing, and interpretation. Our methodological approach generated 49 georeferenced high-resolution TIR and optical orthomosaicked imagery sets. The imagery sets allowed us to identify river-length longitudinal patterns of temperature and to detect, characterize, and classify 260 CWPs. Both stream and CWPs temperatures increased but presented considerable variability with downstream distance. CWPs were non-uniformly distributed along the riverscape, with emergent hyporheic water types dominating, followed by deep pools, shading, side channels, and tributaries. We found associations between CWPs and key physical controls including land use, riparian vegetation, longitudinal and lateral CWP location, and CWP area size, illustrating processes acting at multiple spatial scales. This study provides a basis for future works on the thermal associations with physical controls over a riverscape, and it highlights the major challenges and limitations of the use of UAV-based TIR and optical imagery to be used in future applications. In conjunction with studies of thermally linked ecological processes, the predictions of CWPs can help prioritize river restoration measures as effective climate adaptation tools.

UAV-based canopy textures assess changes in forest structure from long-term degradation

Degraded tropical forests dominate agricultural frontiers and their management is becoming an urgent priority. This calls for a better understanding of the different forest cover states and cost-efficient techniques to quantify the impact of degradation on forest structure. Canopy texture analyses based on Very High Spatial Resolution (VHSR) optical imagery provide proxies to assess forest structures but the mechanisms linking them with degradation have rarely been investigated. To address this gap, we used a lightweight Unmanned Aerial Vehicle (UAV) to map 739 ha of degraded forests and acquire both canopy VHSR images and height model. Thirty-three years of degradation history from Landsat archives allowed us to sample 40 plots in undisturbed, logged, over-logged and burned and regrowth forests in tropical forested landscapes (Paragominas, Pará, Brazil). Fourier (FOTO) and lacunarity textures were used to assess forest canopy structure and to build a typology linking degradation history and current states. Texture metrics capture canopy grain, heterogeneity and openness gradients and correlate with forest structure variability (R2 = 0.58). Similar structures share common degradation history and can be discriminated on the basis of canopy texture alone (accuracy = 55%). Over-logging causes a lowering in forest height, which brings homogeneous textures and of finer grain. We identified the major changes in structures due to fire following logging which changes heterogeneous and intermediate grain into coarse textures. Our findings highlight the potential of canopy texture metrics to characterize degraded forests and thus be used as indicators for forest management and degradation mitigation. Inexpensive and agile UAV open promising perspectives at the interface between field inventory and satellite characterization of forest structure using texture metrics.

Supervised wetland classification using high spatial resolution optical, SAR, and LiDAR imagery

Wetlands are among the most valuable natural resources, being highly beneficial to both the environment and humans. Therefore, it is very important to map and monitor wetlands. Although various remote sensing datasets, including optical, synthetic aperture radar (SAR), light detection and ranging (LiDAR) imagery, have been widely applied to classify wetlands, it is still required to discuss the advantages/limitations of each of these datasets and suggest the best remote sensing methodology for wetland mapping. Thus, the Terra Nova National Park, located in Newfoundland, Canada, was initially selected as the study area to develop a supervised classification method along with object-based image analysis. To this end, different remote sensing-based scenarios were investigated using individual optical, SAR, and LiDAR datasets, as well as their various combinations. In addition, for achieving the highest accuracy, the effects of segmentation scales and the tuning parameters of the random forest (RF) classifier were examined. The results showed that a combination of optical, SAR, and LiDAR images with the segmentation scale of 150, the RF depth of 20, and the RF minimum sample number of 5 provided the highest classification accuracy with the overall accuracy of 87.2%. Moreover, based on the results, approximately 21% and 79% of the study area are covered by wetlands and nonwetlands, respectively. The proposed methodology shows an optimum scenario for future wetland classification tasks and can assist stakeholders in the effective management of wetlands and establishment of necessary policies.

Transferability of the Deep Learning Mask R-CNN Model for Automated Mapping of Ice-Wedge Polygons in High-Resolution Satellite and UAV Images

State-of-the-art deep learning technology has been successfully applied to relatively small selected areas of very high spatial resolution (0.15 and 0.25 m) optical aerial imagery acquired by a fixed-wing aircraft to automatically characterize ice-wedge polygons (IWPs) in the Arctic tundra. However, any mapping of IWPs at regional to continental scales requires images acquired on different sensor platforms (particularly satellite) and a refined understanding of the performance stability of the method across sensor platforms through reliable evaluation assessments. In this study, we examined the transferability of a deep learning Mask Region-Based Convolutional Neural Network (R-CNN) model for mapping IWPs in satellite remote sensing imagery (~0.5 m) covering 272 km2 and unmanned aerial vehicle (UAV) (0.02 m) imagery covering 0.32 km2. Multi-spectral images were obtained from the WorldView-2 satellite sensor and pan-sharpened to ~0.5 m, and a 20 mp CMOS sensor camera onboard a UAV, respectively. The training dataset included 25,489 and 6022 manually delineated IWPs from satellite and fixed-wing aircraft aerial imagery near the Arctic Coastal Plain, northern Alaska. Quantitative assessments showed that individual IWPs were correctly detected at up to 72% and 70%, and delineated at up to 73% and 68% F1 score accuracy levels for satellite and UAV images, respectively. Expert-based qualitative assessments showed that IWPs were correctly detected at good (40–60%) and excellent (80–100%) accuracy levels for satellite and UAV images, respectively, and delineated at excellent (80–100%) level for both images. We found that (1) regardless of spatial resolution and spectral bands, the deep learning Mask R-CNN model effectively mapped IWPs in both remote sensing satellite and UAV images; (2) the model achieved a better accuracy in detection with finer image resolution, such as UAV imagery, yet a better accuracy in delineation with coarser image resolution, such as satellite imagery; (3) increasing the number of training data with different resolutions between the training and actual application imagery does not necessarily result in better performance of the Mask R-CNN in IWPs mapping; (4) and overall, the model underestimates the total number of IWPs particularly in terms of disjoint/incomplete IWPs.

Application of Convolutional Neural Network on Lei Bamboo Above-Ground-Biomass (AGB) Estimation Using Worldview-2

Above-ground biomass (AGB) directly relates to the productivity of forests. Precisely, AGB mapping for regional forests based on very high resolution (VHR) imagery is widely needed for evaluation of productivity. However, the diversity of variables and algorithms and the difficulties inherent in high resolution optical imagery make it complex. In this paper, we explored the potentials of the state-of-art algorithm convolutional neural networks (CNNs), which are widely used for its high-level representation, but rarely applied for AGB estimation. Four experiments were carried out to compare the performance of CNNs and other state-of-art Machine Learning (ML) algorithms: (1) performance of CNN using bands, (2) performance of Random Forest (RF), support vector regression (SVR), artificial neural network (ANN) on bands, and vegetation indices (VIs). (3) Performance of RF, SVR, and ANN on gray-level co-occurrence matrices (GLCM), and exploratory spatial data analysis (ESDA), and (4) performance of RF, SVR, and ANN based on all combined data and ESDA+VIs. CNNs reached satisfactory results (with R2 = 0.943) even with limited input variables (i.e., only bands). In comparison, RF and SVR with elaborately designed data obtained slightly better accuracy than CNN. For examples, RF based on GLCM textures reached an R2 of 0.979 and RF based on all combined data reached a close R2 of 0.974. However, the results of ANN were much worse (with the best R2 of 0.885).

The Impact of Urban Renewal on Land Surface Temperature Changes: A Case Study in the Main City of Guangzhou, China

To improve land use efficiency, urban renewal must also consider urban microclimates and heat islands. Existing research has depended on manual interpretation of high-resolution optical satellite imagery to resolve land surface temperature (LST) changes caused by urban renewal; however, the acquired ground time series data tend to be uneven and unique to specific frameworks. The objective of this study was to establish a more general framework to study LST changes caused by urban renewal using multi-source remote sensing data. Specifically, urban renewal areas during 2007–2017 were obtained by integrating Landsat and yearly Phased Array type L-band Synthetic Aperture Radar (PALSAR) images, and LST was retrieved from Landsat thermal infrared data using the generalized single-channel algorithm. Our results showed that urban renewal land (URL) area accounted for 1.88% of urban land area. Relative LST between URL and general urban land (GUL) of Liwan, Yuexiu, Haizhu, and Tianhe districts dropped by 0.88, 0.42, 0.43, and 0.10 K, respectively, whereas those of Baiyun, Huangpu, Panyu, and Luogang districts presented opposite characteristics, with a rise in the LST of 0.98, 1.03, 1.63, and 2.11 K, respectively. These results are attributable to population density, building density, and landscape pattern changes during the urban renewal process.