High-resolution Aerial Photographs Research Articles

UAV (Unmanned Aerial Vehicle) for Landslide Analysis Case Study in Grenggeng Village, Kebumen District, Central Java

Remote sensing technology has developed rapidly; one of them is data acquisition techniques using UAV (Unmanned Aerial Vehicle). With high-resolution aerial photographs, an unmanned aerial vehicle (UAV) can be a flexible, cost-effective, and accurate monitoring of landslide technique. This research aimed to determine and test the utilization of unmanned aerial vehicles (UAVs) in congested areas. Data was collected at Grenggeng Village, Kebumen Regency, using unmanned aerial vehicles cruising altitude of 90 – 110 meters above ground level and a spatial resolution of 5 – 10 cm over a 0.200 km2 area. In November 2020, the research site will be a landslide area with similar rock lithology to the Halang Formation’s sandstone and claystone layers. Direct field observations revealed the geological structures involved and the rock lithology that produced the slip field, seepage, and the sorts of vegetation that the community had planted. According to aerial photography data, the relief appears to be a straight-line pattern in the direction of the geological structure, the slope of the layers, and different vegetation. Aerial photography using UAV can also be used to carry out rehabilitation and reconstruction techniques.

Satellite monitoring and visualization of vegetation indices for assessing crop productivity

The article discusses approaches to the use of remote sensing and satellite monitoring tools for assessing the productivity of agricultural crops based on the recognition of high-resolution aerial photographs, followed by numerical calculation of vegetation indices and visualization in cross-platform geoinformation systems.

Validation of the U.S. Geological Survey's Land Change Monitoring, Assessment and Projection (LCMAP) Collection 1.0 annual land cover products 1985–2017

The U.S. Geological Survey Land Change Monitoring, Assessment and Projection (USGS LCMAP) has released a suite of annual land cover and land cover change products for the conterminous United States (CONUS). The accuracy of these products was assessed using an independently collected land cover reference sample dataset produced by analysts interpreting Landsat data, high-resolution aerial photographs, and other ancillary data. The reference sample of nearly 25,000 pixels and the accompanying 33-year time series of annual land cover reference labels allowed for a comprehensive assessment of accuracy of the LCMAP land cover and land cover change products. Overall accuracy (± standard error) for the per-pixel assessment across all years for the eight land cover classes was 82.5% (±0.2%). Overall accuracy was consistent year-to-year within a range of 1.5% but varied regionally with lower accuracy in the eastern United States. User's accuracy (UA) and producer's accuracy (PA) for CONUS ranged from the higher accuracies of Water (UA = 96%, PA = 93%) and Tree Cover (UA = 90%, PA = 83%) to the lower accuracies of Wetland (UA = 69%, PA = 74%) and Barren (UA = 43%, PA = 57%). For a binary change/no change classification, UA of change was 13% (±0.5%) and PA was 16% (±0.6%) for CONUS when agreement was defined as a match by the exact year of change. UA and PA improved to 28% and 34% when agreement was defined as the change being detected by the map and reference data within a ± 2-year window. Change accuracy was higher in the eastern United States compared to the western US. UA was 49% (±0.3) and PA was 54% (±0.3) for the footprint of change (defined as the area experiencing at least one land cover change from 1985 to 2017). For class-specific loss and gain when agreement was defined as an exact year match, UA and PA were generally below 30%, with Tree Cover loss being the most accurately mapped change (UA = 25%, PA = 31%). These accuracy results provide users with information to assess the suitability of LCMAP data and information to guide future research for improving LCMAP products, particularly focusing on the challenges of accurately mapping annual land cover change.

Where to start? A new citizen science, remote sensing approach to map recreational disturbance and other degraded areas for restoration planning

This UN Decade on Ecosystem Restoration highlights the capacity of restoration to mitigate trends in biodiversity loss and land degradation. However, many managers lack the tools they need to systematically and comprehensively identify degraded sites to prioritize restoration efforts given limited resources. We developed a novel, inexpensive, low‐tech approach for training and engaging citizen scientists to identify recreational impacts and other degraded areas within a defined unforested area. The mapping process follows four phases: (1) Landscape scans by citizen scientists using Google Earth Pro imagery; (2) A second scan of all marked sites based on high resolution aerial photography; (3) Compilation of basic information about the degraded sites; (4) Addition of associated soil type and plant communities. In the 12,375 ha McDowell Sonoran Preserve (Scottsdale, Arizona), we detected 67 new sites not previously identified by land managers, using an estimated 305 citizen scientist hours and only 30 staff hours. Each site has accompanying information including distance from nearest access point, cause of degradation, and plant and soils detail. After completion, we conducted independent field visits of 33% of the detected sites and verified degradation in all cases. We found that the remotely sensed approach provided better perspective to accurately measure the scale and original source of degradation compared with field visits. The approach can be conducted over a short period of time using citizen scientists, allows managers to undertake landscape level restoration prioritization and planning, and, if repeated, can be used to monitor changes in degradation and restoration over time.

Evaluation of residential structures not covered by aerial photographs used to generate a sampling frame – Nueva Santa Rosa, Guatemala

Background Aerial images are being used more often to map residential structures on the ground in a study area (the sample frame). However, non-coverage bias associated with overhead imagery has not been fully explored. Non-coverage occurs when residential structures are not included in a particular sampling frame. Our study aimed to evaluate non-coverage bias and sensitivity of an aerial photograph methodology in Nueva Santa Rosa, Guatemala, which was used to generate the sampling frame for a larger cross-sectional survey of sanitation, disease, and water quality. Methods High-resolution aerial photographs of Nueva Santa Rosa were overlaid with a grid, and roof images were geo-located within randomly sampled cells, dichotomized by population as very high-density (VHD) or non-VHD. Roofs found on-site were compared to roofs found in photographs to evaluate the numbers and sizes of residences excluded from the sampling frame. Non-coverage proportions were estimated, and sensitivity and specificity were assessed. Results There was no statistically significant difference (1.2%; 95% confidence interval, CI= -12.1-14.6) in non-coverage proportion between VHD segments (39.6%) and non-VHD cells (38.4%). Roof-size range sensitivity and specificity were 66.4% (95% CI=57.6–74.2) and 69.4% (95% CI=54.4–81.3). Conclusions Approximately one-third of residential roofs were missed, perhaps due to outdated photographs. No substantial bias concerning population density appeared to influence our sampling frame. Further assessment of non-coverage bias, possibly expanding the roof size range to modify sensitivity and specificity, should be performed to generate geographically based best practices for overhead-image use.

Coseismic deformation of the ground during large-slip strike-slip ruptures: Finite evolution of “mole tracks”

Abstract To evaluate ground deformation resulting from large (~10 m) coseismic strike-slip displacements, we focus on deformation of the Kekerengu fault during the November 2016 Mw 7.8 Kaikōura earthquake in New Zealand. Combining post-earthquake field observations with analysis of high-resolution aerial photography and topographic models, we describe the structural geology and geomorphology of the rupture zone. During the earthquake, fissured pressure bulges (“mole tracks”) initiated at stepovers between synthetic Riedel (R) faults. As slip accumulated, near-surface “rafts” of cohesive clay-rich sediment, bounded by R faults and capped by grassy turf, rotated about a vertical axis and were internally shortened, thus amplifying the bulges. The bulges are flanked by low-angle contractional faults that emplace the shortened mass of detached sediment outward over less-deformed ground. As slip accrued, turf rafts fragmented into blocks bounded by short secondary fractures striking at a high angle to the main fault trace that we interpret to have originated as antithetic Riedel (R′) faults. Eventually these blocks were dispersed into strongly sheared earth and variably rotated. Along the fault, clockwise rotation of these turf rafts within the rupture zone averaged ~20°–30°, accommodating a finite shear strain of 1.0–1.5 and a distributed strike slip of ~3–4 m. On strike-slip parts of the fault, internal shortening of the rafts averaged 1–2 m parallel to the R faults and ~1 m perpendicular to the main fault trace. Driven by distortional rotation, this contraction of the rafts exceeds the magnitude of fault heave. Turf rafts on slightly transtensional segments of the fault were also bulged and shortened—relationships that can be explained by a kinematic model involving “deformable slats.” In a paleoseismic trench cut perpendicular the fault, one would observe fissures, low-angle thrusts, and steeply dipping strike-slip faults—some cross-cutting one another—yet all may have formed during a single earthquake featuring a large strike-slip displacement.

UAV photogrammetry-based remote sensing and preliminary assessment of the behavior of a landslide in Guizhou, China

Unmanned Aerial Vehicle (UAV) technique has been widely utilized in geohazards assessment. In this work, the effectiveness of the UAV photogrammetry in the remote sensing and assessment of the landslide behavior is demonstrated through a case study of a landslide that occurred in Guizhou, China on 10 June 2018. The post-landslide assessments were conducted through a field investigation by a team of experts; and, two UAV photogrammetry-based surveys were carried out. On the basis of a detailed inspection of the high-resolution aerial photographs collected from the UAV photogrammetry and historical satellite imagery, subsurface stratigraphic configuration revealed from borehole explorations and the local rainfall data collected, the failure mechanism of this landslide is investigated. The occurrence of this landslide is probably attributed to the combined influence of the long-term rainfall and the engineering activities at this site. An automatic landslide cracks recognition model, which is based on the deep learning-based image recognition technique of RetinaNet, is further developed to map the cracks at the landslide site. The effectiveness of this automatic crack recognition model is validated through quantitative comparisons between the landslide cracks recognized and the field survey results. Based upon the landslide cracks identified on two different dates after this landslide event (by the field survey and automatic crack recognition model) and the subsurface displacement revealed from a drilled hole, the evolution behavior of this landslide is analyzed. The results show that the stability of this landslide was not achieved during the first slide in June 2018 and the limit of the landslide was increased much from June to November 2018. In such a situation, the elements at risk in the zones that are potentially impacted by this landslide are identified (with the aid of the UAV images collected), and a preliminary consequence assessment is conducted.

Application and Evaluation of a Deep Learning Architecture to Urban Tree Canopy Mapping

Urban forest is a dynamic urban ecosystem that provides critical benefits to urban residents and the environment. Accurate mapping of urban forest plays an important role in greenspace management. In this study, we apply a deep learning model, the U-net, to urban tree canopy mapping using high-resolution aerial photographs. We evaluate the feasibility and effectiveness of the U-net in tree canopy mapping through experiments at four spatial scales—16 cm, 32 cm, 50 cm, and 100 cm. The overall performance of all approaches is validated on the ISPRS Vaihingen 2D Semantic Labeling dataset using four quantitative metrics, Dice, Intersection over Union, Overall Accuracy, and Kappa Coefficient. Two evaluations are performed to assess the model performance. Experimental results show that the U-net with the 32-cm input images perform the best with an overall accuracy of 0.9914 and an Intersection over Union of 0.9638. The U-net achieves the state-of-the-art overall performance in comparison with object-based image analysis approach and other deep learning frameworks. The outstanding performance of the U-net indicates a possibility of applying it to urban tree segmentation at a wide range of spatial scales. The U-net accurately recognizes and delineates tree canopy for different land cover features and has great potential to be adopted as an effective tool for high-resolution land cover mapping.

Monitoring shoreline shift under rapid sea-level change on the Caspian Sea observed over 60 years of satellite and aerial photo records

The Caspian Sea is the world’s largest inland water body and endures rapid fluctuations in sea level and shoreline changes. The present study investigates and analyzes shoreline changes in the two areas of Tajan and Sefidrud on the southern coasts of the Caspian Sea using aerial photos, affixed with TM, ETM, and OLI images from Landsat satellites, as well as images rendered by Google Earth, all within the time period from 1955 to 2018. Shoreline changes were analyzed using the waterline algorithm on high resolution aerial photos and images from the Google Earth engine. Shoreline extraction on Landsat images was conducted after comparing six different indices, of which the wetness index exhibited the least mean error (1.3 m). Calculations on shoreline changes were undertaken in 4 time periods using the DSAS extension in Arc GIS 10.3 including 1955–1977, 1977–1995, 1995–2002 and 2002–2018. Coastal accretion on the Tajan shoreline after decreases in sea level were evident in the first, third, and fourth time periods, while coastal erosion was discernible in the second time period during significant increases in sea level. The situation, however, was different for the coasts of Sefidrud. The shoreline was divided into 14 distinct parts where the shoreline change rate was inconsistent with the actual changes in sea level. The dismatch between sea level and shoreline changes in the Sefidrud region points to the existence of a different set of factors, in addition to the sea level, which significantly affect the status of the coastlines.

Revealing geomorphic changes after the occurrence of a seismic event on the south coast of Puerto Rico (2020)

A coastal geomorphological study was conducted to assess the impact of seismic events on the southern coast of Puerto Rico (PR). The goal of this study was to provide new information related to the state of the coast and its geomorphic changes associated with the seismic sequence of 28 December 2019 and the 6.4-magnitude earthquake of 7 January 2020. Coastal geomorphic changes and damage to infrastructure were documented using a combination of a geographic information system (GIS), remote sensing techniques (e.g. unmanned aircraft system [UAS] images), ground-penetrating radar (GPR), high-resolution aerial photos, and field inspection techniques. A geospatial analysis was performed to evaluate the effect of the earthquake events and understand its associated effect on coastal communities’ exposure to this hazard. The findings indicate that the 6.4-magnitude earthquake caused remarkable geomorphic changes, generating a new shape for the southern coast of PR. Coastal changes were characterized by localized coastal subsidence, coastal landslides (e.g. rockslides and rockfalls), the collapse of natural arch-forming sea stacks, changes in cliff profiles, lateral spreading, the destruction of coastal infrastructure, inland shoreline migration, and coastal flooding. This information will be useful for designing new management tools and recontextualizing existing coastal planning instruments in post-earthquake scenarios. These findings clarify the importance of seismic events and their influence on coasts within active tectonic settings.

Resilience of trees and the vulnerability of grasslands to climate change in temperate Australian wetlands

Observations from wetlands across the globe suggest a consistent pattern of woody encroachment into wetland grasslands, altering habitat structure and ecological function. The extent to which hydrological changes have contributed to woody invasion of wetland grasslands is unclear. Our objective was to compare rates of woody encroachment in Australian floodplain wetlands between wet and dry hydrological phases. We test the hypothesis that contraction of non-woody wetland vegetation (grasses and rushes) would be concentrated in dry phases, co-incident with recruitment of the River Red Gum Eucalyptus camaldulensis lower in the floodplain. We conduct the first detailed mapping of habitat change in two of the largest forested wetlands in inland Australia, comparing wet and dry hydrological phases. Detailed photogrammetry, supported by extensive ground survey, allowed the interpretation of high resolution aerial photography to vegetation community level. We found a consistent pattern of decline in non-woody vegetation, particularly amongst grasses utilising the C4 photosynthetic pathway. The C4 grasses Pseudoraphis spinescens and Paspalum distichum showed steep declines in the Barmah Millewa and Macquarie Marshes respectively, being replaced by River Red Gum E. camaldulensis. C3 sedges proved more resilient in both systems. Our results suggest that a pattern of tree expansion into non-woody wetland vegetation, characteristic of wetlands across the globe, is a major habitat structural change in the Australian floodplain wetlands studied. Projected hydrological impacts of climate change are likely to further restrict wetland grass foraging habitat in these semi-arid floodplain wetlands.

Winter atmospheric boundary layer observations over sea ice in the coastal zone of the Bay of Bothnia (Baltic Sea)

Abstract. The Hailuoto Atmospheric Observations over Sea ice (HAOS) campaign took place at the westernmost point of Hailuoto island (Finland) between 27 February and 2 March 2020. The aim of the campaign was to obtain atmospheric boundary layer (ABL) observations over seasonal sea ice in the Bay of Bothnia. Throughout 4 d, both fixed-wing and quad-propeller rotorcraft unmanned aerial vehicles (UAVs) were deployed over the sea ice to measure the properties of the lower ABL and to obtain accompanying high-resolution aerial photographs of the underlying ice surface. Additionally, a 3D sonic anemometer, an automatic weather station, and a Halo Doppler lidar were installed on the shore to collect meteorological observations. During the UAV flights, measurements of temperature, relative humidity, and atmospheric pressure were collected at four different altitudes between 25 and 100 m over an area of ∼ 1.5 km2 of sea ice, located 1.1–1.3 km off the shore of Hailuoto's Marjaniemi pier, together with orthomosaic maps of the ice surface below. Altogether the obtained dataset consists of 27 meteorological flights, four photogrammetry missions, and continuous measurements of atmospheric properties from ground-based stations located at the coast. The acquired observations have been quality controlled and post-processed and are available through the PANGAEA repository (https://doi.org/10.1594/PANGAEA.918823, Wenta et al., 2020). The obtained dataset provides us with valuable information about ABL properties over thin, newly formed sea ice cover and about physical processes at the interface of sea ice and atmosphere which may be used for the validation and further improvement of numerical weather prediction (NWP) models.

Semi-Supervised Perception Augmentation for Aerial Photo Topologies Understanding.

Intelligently understanding the sophisticated topological structures from aerial photographs is a useful technique in aerial image analysis. Conventional methods cannot fulfill this task due to the following challenges: 1) the topology number of an aerial photo increases exponentially with the topology size, which requires a fine-grained visual descriptor to discriminatively represent each topology; 2) identifying visually/semantically salient topologies within each aerial photo in a weakly-labeled context, owing to the unaffordable human resources required for pixel-level annotation; and 3) designing a cross-domain knowledge transferal module to augment aerial photo perception, since multi-resolution aerial photos are taken asynchronistically in practice. To handle the above problems, we propose a unified framework to understand aerial photo topologies, focusing on representing each aerial photo by a set of visually/semantically salient topologies based on human visual perception and further employing them for visual categorization. Specifically, we first extract multiple atomic regions from each aerial photo, and thereby graphlets are built to capture the each aerial photo topologically. Then, a weakly-supervised ranking algorithm selects a few semantically salient graphlets by seamlessly encoding multiple image-level attributes. Toward a visualizable and perception-aware framework, we construct gaze shifting path (GSP) by linking the top-ranking graphlets. Finally, we derive the deep GSP representation, and formulate a semi-supervised and cross-domain SVM to partition each aerial photo into multiple categories. The SVM utilizes the global composition from low-resolution counterparts to enhance the deep GSP features from high-resolution aerial photos which are partially-annotated. Extensive visualization results and categorization performance comparisons have demonstrated the competitiveness of our approach.

An object based image analysis applied for volcanic and glacial landforms mapping in Sahand Mountain, Iran

Landform mapping has increasingly become part of the digital domain. While the majority of approaches evaluates Digital Elevation Models (DEM) on a per-pixel basis, some examples exist were object-based image analysis (OBIA) has been applied to terrain data to identify a variety of landforms, including glacial landforms. The main objective of this study is to develop a semi-automated object-based rule set for detecting and delineating volcanic and glacier landforms in the area of the Sahand Mountain, Northern Iran. First, we applied a multi-resolution segmentation algorithm on a freely available Sentinel-2 optical satellite image and then selected the relevant features to define appropriate segmentation scales for each landform category. Object-based rule sets were then developed using spatial (DEM and its derivatives, e.g. slope, aspect, curvature and flow accumulation) and spectral information. Volcanic and glacial landforms were detected and classified into eight classes: caldera, volcanic cone, tuff formation, andesite lava, dacite lava, glacier valley, suspension valley, glacier cirque. An accuracy assessment was applied based on the fuzzy synthetic evaluation technique, together with the error matrix and kappa coefficient, using field data and geomorphological units derived from geological maps and very high resolution aerial photographs. The resulting overall accuracies for each class were 96.2%, 93.3%, 92.4%, 94.2%, 93.01, 95.1, 90.1 and 90.5, respectively. Our research demonstrated that spatial (e.g. density, shape index, length/width) and spectral (e.g. mean, brightness and standard deviation) algorithms together with a grey-level co-occurrence matrix (GLCMs) were efficient features for detecting and mapping volcanic and glacial landforms. We conclude that the OBIA-based algorithms and features provide a high potential for detecting and classifying landforms. Results of this study are of great importance for geomorphology and geology as well as geo-tourism societies and the semi-automated landform mapping contributes to the framework of GIScience.

Mapping the fracture network in the Lilstock pavement, Bristol Channel, UK: manual versus automatic

Abstract. The 100 000 m2 wave-cut pavement in the Bristol Channel near Lilstock, UK, is a world-class outcrop, perfectly exposing a very large fracture network in several thin limestone layers. We present an analysis based on manual interpretation of fracture generations in selected domains and compare it with automated fracture tracing. Our dataset of high-resolution aerial photographs of the complete outcrop was acquired by an unmanned aerial vehicle, using a survey altitude optimized to resolve all fractures. We map fractures and identify fracture generations based on abutting and overprinting criteria, and we present the fracture networks of five selected representative domains. Each domain is also mapped automatically using ridge detection based on the complex shearlet transform method. The automatic fracture detection technique provides results close to the manually traced fracture networks in shorter time but with a bias towards closely spaced Y over X nodes. The assignment of fractures into generations cannot yet be done automatically, because the fracture traces extracted by the automatic method are segmented at the nodes, unlike the manual interpretation in which fractures are traced as a path from fracture tip to fracture tip and consist of several connected segments. This segmentation makes an interpretation of relative age impossible, because the identification of correct abutting relationships requires the investigation of the complete fracture trace by following a clearly defined set of rules. Generations 1 and 2 are long fractures that traverse all domains. Generation 3 is only present in the southwestern domains. Generation 4 follows an ENE–WSW striking trend, is suborthogonal to generations 1 and 2, and abuts on them and generation 3, if present. Generations 5 is the youngest fracture set with a range of orientations, creating polygonal patterns by abutting at all other fracture generations. Our mapping results show that the northeastern domains only contain four fracture generations; thus, the five generations of the outcrop identified in the southwestern domains are either not all present in each of the five domains or vary locally in their geometry, preventing the interpreter from linking the fractures to their respective generation over several spatially separate mapping domains. Fracture intensities differ between domains where the lowest is in the NE with 7.3 m−1 and the highest is in the SW with 10 m−1, coinciding with different fracture orientations and distributions of abutting relationships. Each domain has slightly different fracture network characteristics, and greater connectivity occurs where the development of later shorter fractures is not affected by the stress shadowing of pre-existing longer fractures.

Winter atmospheric boundary layer observations over sea ice in the coastal zone of the Bothnian Bay (Baltic Sea)

Abstract. The Hailuoto Atmospheric Observations over Sea ice (HAOS) campaign took place at the westernmost point of Hailuoto island (Finland) between 27 February and 2 March 2020. The aim of the campaign was to obtain atmospheric boundary layer (ABL) observations over seasonal sea ice in the Bay of Bothnia. Throughout 4 days both fixed-wing and quad-propeller rotorcraft unmanned aerial vehicles (UAV) were deployed over the sea ice to measure the properties of the lower ABL and to obtain accompanying high-resolution aerial photographs of the underlying ice surface. Additionally, a 3D sonic anenometer, an automatic weather station and a Halo Doppler lidar were installed on the shore to collect meteorological observations. During the UAV flights, measurements of temperature, relative humidity and atmospheric pressure were collected at 4 different altitudes between 25 m and 100 m, over an area of ~ 1.5 km2 of sea ice, located 1.1–1.3 km offshore from the Hailuoto Marjaniemi pier, together with orthomosaic maps of the ice surface below. Altogether the obtained dataset consists of 27 meteorological flights, 4 photogrammetry missions and continuous measurements of atmospheric properties from ground-based stations located at the coast. The acquired observations have been quality controlled and post processed and are available through the PANGAEA repository ( https://doi.pangaea.de/10.1594/PANGAEA.918823 , Wenta et al., 2020).

A NEW METHOD OF FAST REGISTRATION OF UNMANNED AERIAL VEHICLE REMOTE SENSING IMAGES BASED-ON AN IMPROVED SURF ALGORITHM

Abstract. Remote sensing system fitted on UAV (Unmanned Aerial Vehicle) can obtain clear images and high-resolution aerial photographs. It has advantages of flexibility, convenience and ability to work full-time. However, there are some problems of UAV image such as small coverage area, large number, irregular overlap, etc. How to obtain a large regional map quickly becomes a major obstacle to UAV remote sensing application. In this paper, a new method of fast registration of UAV remote sensing images was proposed to meet the needs of practical application. This paper used Progressive Sample Consensus (PROSAC) algorithm to improve the matching accuracy by removed a large number of mismatching point pairs of remote sensing image registration based-on SURF (Speed Up Robust Feature) algorithm, and GPU (Graphic Processing Unit) was also used to accelerate the speed of improved SURF algorithm. Finally, geometric verification was used to achieve mosaic accuracy in survey area. The number of feature points obtained by using improved SURF based-on PROSAC algorithm was only 9.5% than that of SURF algorithm. Moreover, the accuracy rate of improved method was about 99.7%, while the accuracy rate of improved SURF algorithm was increased by 8% than SURF algorithm. Moreover, the improved running time of SURFGPU algorithm for UAV remote sensing image registration was a speed of around 16 times than SURF algorithm, and the image matching time had reached millisecond level. Thus, improved SURF algorithm had better matching accuracy and executing speed to meet the requirements of real-time and robustness in UAV remote sensing image registration.

Reliability of using high-resolution aerial photography (red, green and blue bands) for detecting available soil water in agricultural land

The need for irrigation water is influenced by soil water content or more precisely by available water (pF 2.5 and pF 4.2). There is a need for technological breakthroughs in using Unmanned Aerial Vehicle (UAV) to identify water content quickly and broadly and accurately. The study was conducted in an area of ±18 hectares in the Sisim Sub Watershed in September 2019 at 09.00 a.m. Aerial photographs were taken at an altitude of 100 m with DJI Phantom Pro 3.0. The number of observation points was 75 points, where 15 points for validation were calculated based on the map scale. Photo processing was made using Agisoft. The Digital Elevation Model (DEMNAS) with 8.2 m resolution was used to compare the red, green and blue bands. The analysis used was Co-Kriging Geo Statistics Analysis, the compilation of algorithms based on the regression equation and ten index formulations. Validation was done by correlation continued with the regression or paired t-test if the parameter relationship was close. The available water measured in the field ranged from 5.16-48.28%. The results showed that the formulation of soil water content could be run on the Red, Green, and Blue bands, Intensity index, TGI index, ExGreen index and DEMNAS with a weak correlation (below 0.5), where TGI had the highest value (r=0.32). A test of t-pairing was not done because of a weak correlation. The highest estimation of pF 4.2 is DEMNAS (r=0.35), and pF 2.5 was on the TGI index (r=0.4).

Geomorphological processes and landforms of glacier forelands in the upper Aktru River basin (Gornyi Altai), Russia: evidence for rapid recent retreat and paraglacial adjustment

The glaciers in the Aktru River basin of Gornyi Altai, Russia currently represent some of the fastest receding glaciers in the world. Formation of the morainic complexes closest to the contemporary glaciers in the Aktru River basin took place during the 17th - 18th centuries with recession commencing at the end of the 18th century. Coupled with this glacial retreat, earth surface processes and vegetation succession are responding to shape the glacier forelands. This article presents the first geomorphological maps for the upper reaches of the Aktru River basin and focuses on the geomorphological landforms that occur in the rapidly changing glacier forelands. Geomorphological mapping is difficult in steep mountainous regions and, thus, mapping was completed using satellite imagery, field mapping and observations coupled with highresolution aerial photography obtained from Unmanned Aerial Vehicles (UAVs). Critical steps of the procedure used to process UAV imagery and difficulties encountered in this mountainous terrain are noted. The acquired spatial data enable the mapping and classification of small-scale transient geomorphological features such as talus, glacial and glaciofluvial landforms. Their dynamics provide insights into supraglacial and subglacial processes of the glaciers of the Aktru River basin and subsequent paraglacial adjustment. The presented highresolution spatial data, which can also be obtained at high temporal resolutions in the future, can act as a reference frame for geomorphologists and ecologists studying the temporal evolution of glacier forelands of the Aktru River basin during paraglacial adjustment and subsequent colonisation and stabilisation by biota.

Detecting Banana Plantations in the Wet Tropics, Australia, Using Aerial Photography and U-Net

Bananas are the world’s most popular fruit and an important staple food source. Recent outbreaks of Panama TR4 disease are threatening the global banana industry, which is worth an estimated $8 billion. Current methods to map land uses are time- and resource-intensive and result in delays in the timely release of data. We have used existing land use mapping to train a U-Net neural network to detect banana plantations in the Wet Tropics of Queensland, Australia, using high-resolution aerial photography. Accuracy assessments, based on a stratified random sample of points, revealed the classification achieves a user’s accuracy of 98% and a producer’s accuracy of 96%. This is more accurate compared to existing (manual) methods, which achieved a user’s and producer’s accuracy of 86% and 92% respectively. Using a neural network is substantially more efficient than manual methods and can inform a more rapid respond to existing and new biosecurity threats. The method is robust and repeatable and has potential for mapping other commodities and land uses which is the focus of future work.