Unmanned Aircraft Systems Imagery Research Articles

Impacts of Aquatic Vegetation Dynamics on Nitrate Removal in Karst Agricultural Streams: Insights from Unmanned Aircraft Systems and In Situ Sensing

Highlights Aerial image results highlight the spatial and temporal variability of duckweed coverage in the headwater stream. In situ sensors suggest algae has a shortened season compared to other agricultural streams. Diurnal nitrate variability changed with shifts in aquatic vegetation and dissolved oxygen. Coupling in situ and aerial results elucidated reasons for limitations in nitrate concentration predictions. Abstract. Fate and transport of nutrients in karst streams remains a pressing research need. The objective of this study was to couple high-frequency and remote imagery data to quantify spatial and temporal variability of aquatic vegetation and determine the associated impacts on in-stream nitrate removal in karst headwater streams. The study was conducted in a spring-fed karst stream in the Inner-Bluegrass region of Central Kentucky, USA. Ten Unmanned Aircraft System (UAS) campaigns were coupled with three-years of high frequency in situ data of water quality. Automated segmentation and image classification analysis was performed on UAS imagery, and spatial variability in floating aquatic macrophytes was quantified. Results demonstrated the utility of UAS images to capture spatiotemporal variability of floating aquatic macrophytes (duckweed) with high accuracy, but the UAS analysis poorly predicted algal biomass dynamics due to spectral interferences in the water column and shading by duckweed. Further, in situ water quality data was used to estimate stream metabolism using the Bayesian Single Station Estimator (BASE) model, with results demonstrating primary production and ecosystem respiration was driven by algal biomass. Stream metabolism displayed a shorter season (March-August) relative to other agricultural streams in the region (March-October), likely reflecting the hydraulic structures in the channel which reduce flow velocities in the stream reach and promote transition to duckweed cover during the summer. The impact of aquatic vegetation dynamics on nitrate was assessed using diurnal analysis and regression with dissolved oxygen. Results for March through November showed an average daily diurnal variation of 0.25 mgN/L, which was more than two-fold greater than winter months. During the growing season, maxima generally occurred between 6 a.m.-2 p.m., and minima occurred between 2 p.m.-12 a.m., but varied depending on prominent aquatic vegetation and dissolved oxygen dynamics. These results suggest shifting N removal mechanisms as aquatic vegetation changes throughout the year. Implications for numerical modeling of fluvial nitrate fluxes are illustrated through the evaluation of a previously developed AI model simulating nitrate concentrations at the watershed outlet. The findings highlight the importance of integrating novel datasets, such as those presented in this study, to evaluate and improve predictions of numerical models. Keywords: Aquatic vegetation, Water quality sensing, Karst agroecosystem, Nitrate, watershed.

Aerial characterization of surface depressions in urban watersheds

Digital elevation models (DEM) are one of the most fundamental inputs for hydrological modeling. It has been a common practice to remove all surface depressions in a DEM as they are assumed to be data errors. The emerging technology of unmanned aircraft systems (UAS) provides an opportunity to re-examine this assumption at the hyperspatial resolution. This study was the first attempt to characterize small surface depressions in urban environments using UAS imagery. Using an urban area in south Texas as the study site, UAS flights were conducted to yield hybrid DEMs at the resolution of 8–14 cm, coupled with comprehensive ground truth collection. Surface depressions identified from the UAS DEMs were first corrected based on the vertical accuracy of DEMs and then validated through field surveys, with comparisons to two existing LiDAR DEMs (1-m and 10-m). The hydrological impacts of different DEM-derived estimates of catchment depression storage were examined using the Curve Number method across different design storms. Results show that the UAS DEMs outperformed the LiDAR DEMs in describing the microtopographic control of urban overland flow and associated hydrological connectivity across built and natural features. The 8-cm UAS DEM revealed 926% more depression storage than the 10-m LiDAR DEM. This demonstrates a compelling correlation between increasing DEM resolution and enhanced quantification of depression volume. Consequently, the increased depression storage reduced surface runoff by 41% under a two-year design storm and 13% under a 200-year design storm. The results suggest a strong relationship between the DEM resolution and the derived depression estimates, aligning with the fractal nature of watershed systems. Also, the results indicate that the centimeter-level UAS DEMs were not immune from problems. They could yield fake depressions caused by factors such as vegetation, temporary street objects, and underground sewer pipes. The findings of this study suggest the need to quantify the relationships between DEM resolution and associated hydrological attributes and develop new digital drainage analysis algorithms that could effectively incorporate UAS data into urban hydrological modeling.

Object Recognition of a GCP Design in UAS Imagery Using Deep Learning and Image Processing—Proof of Concept Study

Image-based unmanned aircraft systems (UASs) are used in a variety of geodetic applications. Precise 3D terrain surface mapping requires ground control points (GCPs) for scaling and (indirect) georeferencing. In image analysis software (e.g., Agisoft Metashape), the images can be generated to a 3D point cloud using Structure-from-Motion (SfM). In general, the conventional GCP design for UAS flights is a checkerboard pattern, which is provided in the software and used for automatic marker detection in each image. When changing the pattern, manual work would be required by picking the GCP individually by hand. To increase the level of automation in the evaluation, this article aims to present a workflow that automatically detects a new edge-based GCP design pattern in the images, calculates their center points, and provides this information to the SfM software. Using the proposed workflow based on deep learning (DL) and image processing, the quality of the resulting 3D model can be equated to the result with GCP center points picked by human evaluator. Consequently, the workload can be accelerated with this approach.

An Improved Deep Learning Approach for Retrieving Outfalls Into Rivers From UAS Imagery

Outfalls into rivers are the final gate of anthropogenic pollution flowing to receiving waters, which means that outfall surveys are significant to basin environmental protection and ecosystem health management. Unmanned aircraft systems (UAS) with high spatial resolution imagery have become important data for ongoing surveys of outfalls. However, outfalls retrieval from UAS imagery is inefficient to visual interpretation and a challenging task for traditional spectral-based and object-oriented classification methods given the problems of salt-and-pepper noise and scale selection. In this study, an improved geo-deep learning approach based on the faster region convolutional neural network (R-CNN) architecture (GDCNN-outfalls) is proposed for retrieving outfalls into rivers with UAS imagery. In the proposed method, three tactics—anchor size, region of interest (RoI), and hard negative mining—were adopted to optimize the benchmark Faster R-CNN application in outfalls retrieval. Meanwhile, a geo-classifier module with digital surface model (DSM) enhancement and a spatial activation function was integrated with the Faster R-CNN architecture to generate GDCNN-outfalls. The validation experiments indicated that GDCNN-outfalls improved the performance of Faster R-CNN in outfall retrieval by suppressing false positive (FPs) from 33.52% to 26.14% and increasing the F1 score from 0.72 to 0.75. The test results confirm the performance of GDCNN-outfalls with a recall of 79.3% and higher precision (48.4%) than that of Faster R-CNN (2.1%), also show the GDCNN-outfalls is ten times faster than visual interpretation. This study demonstrates that the combination of deep learning and UAS techniques can be a feasible solution to detect outfalls in outfall surveys.

Evaluation of Crop Health Status With UAS Multispectral Imagery

This study presents the results of a field experiment conducted for assessing the crop health status of several barley and oat crop fields in Prince Edward Island, Canada. The crop fields were mapped with an Unmanned Aircraft System (UAS) and the crop health status was assessed through the Green Area Index (GAI) and vegetation indices (VIs). GAI maps were produced from the UAS imagery and VIs using machine learning pipelines with several regression algorithms (Multiple Linear Models, Support Vector Machines, Random Forests, and Artificial Neural Networks) along with a feature selection strategy. The Random Forests algorithm was shown to be the best algorithm for GAI prediction with an average relative Root Mean Square Error of 10.86% and a Mean Absolute Error of 0.67. The resulting GAI maps and the regression feature space were classified with Random Forests to discriminate between vigorous and stressed crop areas. We achieved a mean overall accuracy of 94%. The limits of the study are also presented.

MAPPING BARLEY LODGING WITH UAS MULTISPECTRAL IMAGERY AND MACHINE LEARNING

Abstract. Unmanned Aircraft Systems (UAS) are demonstrated cost- and time-effective remote sensing platforms for precision agriculture applications and crop damage monitoring. In this study, lodging damage on barley crops has been mapped from UAS imagery that was acquired over multiple barley fields with extensive lodging damages in two aerial surveys. A Random Forests classification model was trained and tested for the discrimination of lodged barley with an overall accuracy of 99.7% on the validation dataset. The crop areas with lodging were automatically delineated by vector analysis and compared to manually delineated areas using two spatial accuracy metrics, the Area Goodness of Fit (AGoF) and the Boundary Mean Positional Error (BMPE). The average AGoF was 97.95% and the average BMPE was 0.235 m.

Early Detection of Encroaching Woody Juniperus virginiana and Its Classification in Multi-Species Forest Using UAS Imagery and Semantic Segmentation Algorithms

Woody plant encroachment into grasslands ecosystems causes significantly ecological destruction and economic losses. Effective and efficient management largely benefits from accurate and timely detection of encroaching species at an early development stage. Recent advances in unmanned aircraft systems (UAS) enabled easier access to ultra-high spatial resolution images at a centimeter level, together with the latest machine learning based image segmentation algorithms, making it possible to detect small-sized individuals of target species at early development stage and identify them when mixed with other species. However, few studies have investigated the optimal practical spatial resolution of early encroaching species detection. Hence, we investigated the performance of four popular semantic segmentation algorithms (decision tree, DT; random forest, RF; AlexNet; and ResNet) on a multi-species forest classification case with UAS-collected RGB images in original and down-sampled coarser spatial resolutions. The objective of this study was to explore the optimal segmentation algorithm and spatial resolution for eastern redcedar (Juniperus virginiana, ERC) early detection and its classification within a multi-species forest context. To be specific, firstly, we implemented and compared the performance of the four semantic segmentation algorithms with images in the original spatial resolution (0.694 cm). The highest overall accuracy was 0.918 achieved by ResNet with a mean interaction over union at 85.0%. Secondly, we evaluated the performance of ResNet algorithm with images in down-sampled spatial resolutions (1 cm to 5 cm with 0.5 cm interval). When applied on the down-sampled images, ERC segmentation performance decreased with decreasing spatial resolution, especially for those images coarser than 3 cm spatial resolution. The UAS together with the state-of-the-art semantic segmentation algorithms provides a promising tool for early-stage detection and localization of ERC and the development of effective management strategies for mixed-species forest management.

Exploring the Suitability of UAS-Based Multispectral Images for Estimating Soil Organic Carbon: Comparison with Proximal Soil Sensing and Spaceborne Imagery

Soil organic carbon (SOC) is a variable of vital environmental significance in terms of soil quality and function, global food security, and climate change mitigation. Estimation of its content and prediction accuracy on a broader scale remain crucial. Although, spectroscopy under proximal sensing remains one of the best approaches to accurately predict SOC, however, spectroscopy limitation to estimate SOC on a larger spatial scale remains a concern. Therefore, for an efficient quantification of SOC content, faster and less costly techniques are needed, recent studies have suggested the use of remote sensing approaches. The primary aim of this research was to evaluate and compare the capabilities of small Unmanned Aircraft Systems (UAS) for monitoring and estimation of SOC with those obtained from spaceborne (Sentinel-2) and proximal soil sensing (field spectroscopy measurements) on an agricultural field low in SOC content. Nine calculated spectral indices were added to the remote sensing approaches (UAS and Sentinel-2) to enhance their predictive accuracy. Modeling was carried out using various bands/wavelength (UAS (6), Sentinel-2 (9)) and the calculated spectral indices were used as independent variables to generate soil prediction models using five-fold cross-validation built using random forest (RF) and support vector machine regression (SVMR). The correlation regarding SOC and the selected indices and bands/wavelengths was determined prior to the prediction. Our results revealed that the selected spectral indices slightly influenced the output of UAS compared to Sentinel-2 dataset as the latter had only one index correlated with SOC. For prediction, the models built on UAS data had a better accuracy with RF than the two other data used. However, using SVMR, the field spectral prediction models achieved a better overall result for the entire study (log(1/R), RPD = 1.40; R2CV = 0.48; RPIQ = 1.65; RMSEPCV = 0.24), followed by UAS and then Sentinel-2, respectively. This study has shown that UAS imagery can be exploited efficiently using spectral indices.

Unmanned Aircraft System‐ (UAS‐) Based High‐Throughput Phenotyping (HTP) for Tomato Yield Estimation

Yield prediction and variety selection are critical components for assessing production and performance in breeding programs and precision agriculture. Since plants integrate their genetics, surrounding environments, and management conditions, crop phenotypes have been measured over cropping seasons to represent the traits of varieties. These days, UAS (unmanned aircraft system) provides a new opportunity to collect high‐quality images and generate reliable phenotypic data efficiently. Here, we propose high‐throughput phenotyping (HTP) from multitemporal UAS images for tomato yield estimation. UAS‐based RGB and multispectral images were collected weekly and biweekly, respectively. The shape of the features of tomatoes such as canopy cover, canopy, volume, and vegetation indices derived from UAS imagery was estimated throughout the entire season. To extract time‐series features from UAS‐based phenotypic data, crop growth and growth rate curves were fitted using mathematical curves and first derivative equations. Time‐series features such as the maximum growth rate, day at a specific event, and duration were extracted from the fitted curves of different phenotypes. The linear regression model produced high R2 values even with different variable selection methods: all variables (0.79), forward selection (0.7), and backward selection (0.77). With factor analysis, we figured out two significant factors, growth speed and timing, related to high‐yield varieties. Then, five time‐series phenotypes were selected for yield prediction models explaining 65 percent of the variance in the actual harvest. The phenotypic features derived from RGB images played more important roles in prediction yield. This research also demonstrates that it is possible to select lower‐performing tomato varieties successfully. The results from this work may be useful in breeding programs and research farms for selecting high‐yielding and disease‐/pest‐resistant varieties.

Combined geometric-radiometric and neural network approach to shallow bathymetric mapping with UAS imagery

There is a pressing need for shallow bathymetry in regions around the world, including in fragile marine ecosystems, such as coral reefs. Unfortunately, mapping these shallow-water areas using in situ or boat-based methods is expensive, time consuming, and potentially dangerous, due to the need to put personnel and/or boats in high-energy nearshore areas, which may contain rocks, reefs, and other submerged hazards. For this reason, passive remote sensing methods of shallow bathymetric mapping have been rapidly growing in interest and usage. Two general categories of approaches using passive imaging can be distinguished as: (1) those based on geometry (stereo-photogrammetry); (2) those based on radiometry (attenuation of light in the water column). These two broad categories of approaches have been tested on imagery collected using aircraft and satellites in numerous previous studies. Recent work within the geometric category includes processing unmanned aircraft system (UAS) imagery using Structure from Motion and Multi-View Stereo (SfM-MVS) photogrammetry, while radiometric methods encompass the broad range of spectral bathymetry retrieval algorithms. Each broad approach category has advantages and disadvantages. Fortunately, the geometric and radiometric approaches are highly complementary. The geometric approaches tend to work best in areas of high bottom texture, which facilitates feature matching in the SfM-MVS software. Meanwhile, the radiometric approaches work best in relatively homogeneous bottom types. To leverage the strengths of each type of approach and overcome their respective weaknesses, this work develops and tests a combined geometric-radiometric bathymetric mapping approach designed for shallow-water mapping from UAS imagery. Four radiometric models of varying complexity are tested, ranging from a color-based lookup table approach to neural networks. Two UAS flights on Buck Island, off St Croix in the U.S. Virgin Islands (USVI), are used to assess the accuracy of the methodology when compared to aerial bathymetric lidar data. The results show that the combined geometric-radiometric approach provides an increase in spatial coverage of up to 61% and improvement (i.e. decrease) in 95th percentile error of up to 49% when compared to traditional refraction-corrected SfM-MVS bathymetry.

Delineation of Crop Field Areas and Boundaries from UAS Imagery Using PBIA and GEOBIA with Random Forest Classification

Unmanned aircraft systems (UAS) have been proven cost- and time-effective remote-sensing platforms for precision agriculture applications. This study presents a method for automatic delineation of field areas and boundaries that uses UAS multispectral orthomosaics acquired over 7 vegetated fields having a variety of crops in Prince Edward Island (PEI). This information is needed by crop insurance agencies and growers for an accurate determination of crop insurance premiums. The field areas and boundaries were delineated by applying both a pixel-based and an object-based supervised random forest (RF) classifier applied to reflectance and vegetation index images, followed by a vectorization pipeline. Both methodologies performed exceptionally well, resulting in a mean area goodness of fit (AGoF) for the field areas greater than 98% and a mean boundary mean positional error (BMPE) lower than 0.8 m for the seven surveyed fields.

Delineation of Bare Soil Field Areas from Unmanned Aircraft System Imagery with the Mean Shift Unsupervised Clustering and the Random Forest Supervised Classification

The use of aerial remote sensing platforms such as Unmanned Aircraft Systems (UAS) has been proven as a cost and time effective way to perform tasks related to precision agriculture and decision making. Two machine learning (ML) algorithms have been implemented on UAS blue and red band imagery to delineate field areas and extents of various bare soil fields: the Random Forest non-parametric supervised classifier and the unsupervised non-parametric Mean Shift clustering algorithm. Both ML algorithms perform exceptionally well. The mean Area Goodness of Fit (AGoF) for bare soil areas was greater than 99% and the mean Boundary Mean Positional Error (BMPE) was lower than 0.6 m for the 11 surveyed fields. Such precisions with ML algorithms will enable an easy automated field boundary delineation based on UAS imagery. Such information is needed by growers and crop insurance agencies for an accurate determination of the crop insurance premiums.

Plantation Loblolly Pine Seedling Counts with Unmanned Aerial Vehicle Imagery: A Case Study

Abstract Abstract An unmanned aircraft system was evaluated for its potential to capture imagery for use in plantation loblolly pine (Pinus taeda L.) regeneration surveys. Five stands located in the Virginia Piedmont were evaluated. Imagery was collected using a recreational grade unmanned aerial vehicle at three flight heights above ground with a camera capable of capturing red–green–blue imagery. Two computer vision approaches were evaluated for their potential to automatically detect seedlings. The results of the study indicated that the proposed methods were limited in capability of generating reliable counts of seedlings in the locations evaluated. In conditions with low numbers of natural seedlings and sufficiently large planted seedlings, the detection methods performed with higher levels of accuracy. Challenges including global positioning system errors and image distortion made comparisons between ground samples and imagery difficult. In summary, unmanned aircraft systems have potential for use in plantation pine regeneration surveys if the challenges encountered can be addressed. Study Implications: Following the establishment of a pine plantation, it is important to estimate survival and possible recruitment of natural conifers. As the popularity of unmanned aircraft systems (UAS) has increased, forest managers have begun to explore their use for resource assessment. This study investigated using imagery captured with a recreational grade UAS, in conjunction with automated computer vision counting techniques, for use in regeneration surveys. The results of this research indicate that significant challenges must be addressed before UAS can become an integral component of survival assessments. Aircraft constraints, legal restrictions, low image quality, and high levels of natural pine regeneration limited the success of the proposed methods. In selected cases, however, favorable conditions led to accurate detection. Additionally, UAS imagery has the potential for assessing other stand characteristics such as competing vegetation and drainage patterns. Going forward, UAS imagery and automated counting approaches have the potential to supplement, but not fully replace, ground regeneration surveys if the challenges encountered in this study can be addressed.

Site-specific irrigation management in a sub-humid climate using a spatial evapotranspiration model with satellite and airborne imagery

Variable Rate Irrigation (VRI) considers spatial variability in soil and plant characteristics to optimize irrigation management in agricultural fields. The advent of unmanned aircraft systems (UAS) creates an opportunity to utilize high-resolution (spatial and temporal) imagery into irrigation management due to decreasing costs, ease of operation, and reduction of regulatory constraints. This research aimed to evaluate the use of UAS data for VRI, and to quantify the potential of VRI in terms of relative crop and water response. Irrigation treatments were: (1) VRI using Landsat imagery (VRI-L), (2) VRI using UAS imagery (VRI-U), (3) uniform (U), and (4) rainfed (R). An updated remote-sensing-based evapotranspiration and water balance model, incorporating soil water measurements, was used to make prescriptions for the VRI treatments at a field site in eastern Nebraska. In 2017, the mean prescribed seasonal irrigation depth (Ip) for VRI-L was significantly greater (α = 0.05) than the Ip for U for soybean. In 2018, Ip for soybean was greatest for VRI-U treatment followed by the U and VRI-L treatments, with all being significantly different from each other. No significant differences in Ip for maize were observed in 2017 or 2018. In all crop-year combinations, the VRI and U treatments had significantly greater evapotranspiration (ET) than the R treatment. Yield differences among treatments were not significant (except for rainfed maize compared to VRI-L in 2017). For maize in 2017, IWUE for VRI-L was comparable to the U treatment. The UAS imagery was a better match for the scale of crop management than Landsat imagery, particularly for thermal data. The multispectral UAS data was successfully used in the crop coefficient ET model for real-time irrigation, but using UAS to determine accurate canopy temperatures for surface energy balance modeling remains a challenge.

Weed Mapping with UAS Imagery and a Bag of Visual Words Based Image Classifier

Weed detection with aerial images is a great challenge to generate field maps for site-specific plant protection application. The requirements might be met with low altitude flights of unmanned aerial vehicles (UAV), to provide adequate ground resolutions for differentiating even single weeds accurately. The following study proposed and tested an image classifier based on a Bag of Visual Words (BoVW) framework for mapping weed species, using a small unmanned aircraft system (UAS) with a commercial camera on board, at low flying altitudes. The image classifier was trained with support vector machines after building a visual dictionary of local features from many collected UAS images. A window-based processing of the models was used for mapping the weed occurrences in the UAS imagery. The UAS flight campaign was carried out over a weed infested wheat field, and images were acquired between a 1 and 6 m flight altitude. From the UAS images, 25,452 weed plants were annotated on species level, along with wheat and soil as background classes for training and validation of the models. The results showed that the BoVW model allowed the discrimination of single plants with high accuracy for Matricaria recutita L. (88.60%), Papaver rhoeas L. (89.08%), Viola arvensis M. (87.93%), and winter wheat (94.09%), within the generated maps. Regarding site specific weed control, the classified UAS images would enable the selection of the right herbicide based on the distribution of the predicted weed species.

A pilot(less) study on the use of an unmanned aircraft system for studying polar bears (Ursus maritimus)

Unmanned aircraft systems (UAS) are increasingly popular tools for studying wildlife ecology. The non-invasive aspect of UAS and the ability to collect a large amount of high-resolution imagery should be of interest to polar bear (Ursus maritimus) researchers who face logistic challenges with field work and developing minimally invasive methods. We opportunistically observed the behavioural reactions of three adult male polar bears during UAS surveys in the summer of 2016. We recorded vigilance behaviours and compared them to previously published vigilance behaviours during wildlife-viewing activities by Dyck and Baydack (2004). The number of vigilance events was 13.4 ± 3.7 (SE) and vigilance bout lengths was 18.7 ± 2.6 s (SE), which is similar to reported results by Dyck and Baydack (2004). To estimate detection probabilities of polar bears from UAS imagery, we had two independent observers review mosaics and 80% of known bear locations were identified. Our preliminary results suggest that UAS are capable of detecting polar bears using RGB imagery in a relatively non-invasive manner. Before UAS can be integrated into large-scale polar bear studies, further research is required to formally assess behavioural impacts with unhabituated individuals in the wild, and model factors influencing detection probabilities.

Early Stage Forest Windthrow Estimation Based on Unmanned Aircraft System Imagery

Strong wind disturbances can affect large forested areas and often occur irregularly within a forest. Due to this, identifying damaged sites and estimating the extent of these losses are crucial for the harvesting management of salvage logging. Furthermore, the location should be surveyed as soon as possible after the disturbance to prevent the degradation of fallen trees. A fixed-wing type of unmanned aircraft system (UAS) with a compact digital camera was used in this study. The imagery was acquired on approximately 200 hectares where five large windthrow areas had occurred. The objective of the study was to determine the location of the windthrow areas using a semi-automatic approach based on the UAS imagery, and on the combination of UAS imagery with airborne laser scanning (ALS). The results were compared with reference data measured by global navigation satellite system (GNSS) devices. At the same time, windthrow areas were derived from Landsat imagery to investigate whether the UAS imagery would have significantly more accurate results. GNSS measurements and Landsat imagery are currently used in forestry on an operational level. The salvage logging was estimated for each forest stand based on the estimated areas and volume per hectare obtained from the forest management plan. The results from the UAS (25.09 ha) and the combined UAS/ALS (25.56 ha) methods were statistically similar to the reference GNSS measurements (25.39 ha). The result from Landsat, at 19.8 ha, was not statistically similar to the reference GNSS measurements or to the UAS and UAS/ALS methods. The estimate of salvage logging for the whole area, from UAS imagery and the forest management plan, overestimated the actual salvage logging measured by foresters by 4.93% (525 m3), when only the most represented tree species were considered. The UAS/ALS combination improved the preliminary results of determining windthrow areas which lead to decreased editing time for all operators. The UAS imagery shows potential for application to early-stage surveys of windthrow areas in forests. The advantages of this method are that it provides the ability to conduct flights immediately after the disturbance, the foresters do not need to walk within the affected areas which decreases the risk of injury, and allows flights to be conducted on cloudy days. The orthomosaic of the windthrow areas, as a by-product of data processing in combination with forest maps and forest road maps, can be used as a tool to plan salvage logging.

Mapping orangutan habitat and agricultural areas using Landsat OLI imagery augmented with unmanned aircraft system aerial photography

ABSTRACTConservation of the Sumatran orangutans’ (Pongo abelii) habitat is threatened by change in land use/land cover (LULCC), due to the logging of its native primary forest habitat, and the primary forest conversion to oil palm, rubber tree, and coffee plantations. Frequent LULCC monitoring is vital to rapid conservation interventions. Due to the costs of high-resolution satellite imagery, researchers are forced to rely on cost-free sources (e.g. Landsat), those, however, provide images at a moderate-to-low resolution (e.g. 15–250 m), permitting identification only general LULC classes, and limit the detection of small-scale deforestation or degradation. Here, we combine Landsat imagery with very high-resolution imagery obtained from an unmanned aircraft system (UAS). ​The UAS imagery was used as ‘drone truthing’ data to train image classification algorithms. Our results show that UAS data can successfully be used to help discriminate similar land-cover/use classes (oil palm plantation vs. reforestation vs. logged forest) with consistently high identification of over 75% on the generated thematic map, where the oil palm detection rate was as high as 89%. Because UAS is employed increasingly in conservation proWjects, this approach can be used in a large variety of them to improve land-cover classification or aid-specific mapping needs.

Remote Sensing With Simulated Unmanned Aircraft Imagery for Precision Agriculture Applications

An important application of unmanned aircraft systems (UAS) may be remote-sensing for precision agriculture, because of its ability to acquire images with very small pixel sizes from low altitude flights. The objective of this study was to compare information obtained from two different pixel sizes, one about a meter (the size of a small vegetation plot) and one about a millimeter. Cereal rye (Secale cereale) was planted at the Beltsville Agricultural Research Center for a winter cover crop with fall and spring fertilizer applications, which produced differences in biomass and leaf chlorophyll content. UAS imagery was simulated by placing a Fuji IS-Pro UVIR digital camera at 3-m height looking nadir. An external UV-IR cut filter was used to acquire true-color images; an external red cut filter was used to obtain color-infrared-like images with bands at near-infrared, green, and blue wavelengths. Plot-scale Green Normalized Difference Vegetation Index was correlated with dry aboveground biomass (r = 0.58), whereas the Triangular Greenness Index (TGI) was not correlated with chlorophyll content. We used the SamplePoint program to select 100 pixels systematically; we visually identified the cover type and acquired the digital numbers. The number of rye pixels in each image was better correlated with biomass (r = 0.73), and the average TGI from only leaf pixels was negatively correlated with chlorophyll content (r = -0.72). Thus, better information for crop requirements may be obtained using very small pixel sizes, but new algorithms based on computer vision are needed for analysis. It may not be necessary to geospatially register large numbers of photographs with very small pixel sizes. Instead, images could be analyzed as single plots along field transects.