Multispectral Imagery Data Research Articles

Large-Scale River Mapping Using Contrastive Learning and Multi-Source Satellite Imagery

River system is critical for the future sustainability of our planet but is always under the pressure of food, water and energy demands. Recent advances in machine learning bring a great potential for automatic river mapping using satellite imagery. Surface river mapping can provide accurate and timely water extent information that is highly valuable for solid policy and management decisions. However, accurate large-scale river mapping remains challenging given limited labels, spatial heterogeneity and noise in satellite imagery (e.g., clouds and aerosols). In this paper, we propose a new multi-source data-driven method for large-scale river mapping by combining multi-spectral imagery and synthetic aperture radar data. In particular, we build a multi-source data segmentation model, which uses contrastive learning to extract the common information between multiple data sources while also preserving distinct knowledge from each data source. Moreover, we create the first large-scale multi-source river imagery dataset based on Sentinel-1 and Sentinel-2 satellite data, along with 1013 handmade accurate river segmentation mask (which will be released to the public). In this dataset, our method has been shown to produce superior performance (F1-score is 91.53%) over multiple state-of-the-art segmentation algorithms. We also demonstrate the effectiveness of the proposed contrastive learning model in mapping river extent when we have limited and noisy data.

Remote Sensing-Based Statistical Approach for Defining Drained Lake Basins in a Continuous Permafrost Region, North Slope of Alaska

Lake formation and drainage are pervasive phenomena in permafrost regions. Drained lake basins (DLBs) are often the most common landforms in lowland permafrost regions in the Arctic (50% to 75% of the landscape). However, detailed assessments of DLB distribution and abundance are limited. In this study, we present a novel and scalable remote sensing-based approach to identifying DLBs in lowland permafrost regions, using the North Slope of Alaska as a case study. We validated this first North Slope-wide DLB data product against several previously published sub-regional scale datasets and manually classified points. The study area covered >71,000 km2, including a >39,000 km2 area not previously covered in existing DLB datasets. Our approach used Landsat-8 multispectral imagery and ArcticDEM data to derive a pixel-by-pixel statistical assessment of likelihood of DLB occurrence in sub-regions with different permafrost and periglacial landscape conditions, as well as to quantify aerial coverage of DLBs on the North Slope of Alaska. The results were consistent with previously published regional DLB datasets (up to 87% agreement) and showed high agreement with manually classified random points (64.4–95.5% for DLB and 83.2–95.4% for non-DLB areas). Validation of the remote sensing-based statistical approach on the North Slope of Alaska indicated that it may be possible to extend this methodology to conduct a comprehensive assessment of DLBs in pan-Arctic lowland permafrost regions. Better resolution of the spatial distribution of DLBs in lowland permafrost regions is important for quantitative studies on landscape diversity, wildlife habitat, permafrost, hydrology, geotechnical conditions, and high-latitude carbon cycling.

Combination of Sentinel-2 and PALSAR-2 for Local Climate Zone Classification: A Case Study of Nanchang, China

Local climate zone (LCZ) maps have been used widely to study urban structures and urban heat islands. Because remote sensing data enable automated LCZ mapping on a large scale, there is a need to evaluate how well remote sensing resources can produce fine LCZ maps to assess urban thermal environments. In this study, we combined Sentinel-2 multispectral imagery and dual-polarized (HH + HV) PALSAR-2 data to generate LCZ maps of Nanchang, China using a random forest classifier and a grid-cell-based method. We then used the classifier to evaluate the importance scores of different input features (Sentinel-2 bands, PALSAR-2 channels, and textural features) for the classification model and their contribution to each LCZ class. Finally, we investigated the relationship between LCZs and land surface temperatures (LSTs) derived from summer nighttime ASTER thermal imagery by spatial statistical analysis. The highest classification accuracy was 89.96% when all features were used, which highlighted the potential of Sentinel-2 and dual-polarized PALSAR-2 data. The most important input feature was the short-wave infrared-2 band of Sentinel-2. The spectral reflectance was more important than polarimetric and textural features in LCZ classification. PALSAR-2 data were beneficial for several land cover LCZ types when Sentinel-2 and PALSAR-2 were combined. Summer nighttime LSTs in most LCZs differed significantly from each other. Results also demonstrated that grid-cell processing provided more homogeneous LCZ maps than the usual resampling methods. This study provided a promising reference to further improve LCZ classification and quantitative analysis of local climate.

Characterizing building materials using multispectral imagery and LiDAR intensity data

This paper addresses the underlying bottleneck of unknown materials and material characteristics for assessing the life cycle of an existing structure or considering interventions. This is done by classifying and characterizing common building materials with two readily accessible, remote sensing technologies: multispectral imaging and Light Detection and Ranging (LiDAR). A total of 142 samples, including concrete of 3 different water/cement ratios, 2 mortar types, and 2 brick types (each type fired at 3 different temperatures) were scanned using a 5-band multispectral camera in the visible, RedEdge, and Near Infrared range and 2 laser scanners with different wavelengths. A Partial Least Squares Discriminant Analysis model was developed to classify the main materials and then the subcategories within each material type. With multispectral data, an 82.75% average correct classification rate was achieved (improving to 83.07% when combined with LiDAR intensity data), but the effect was not uniformly positive. This paper demonstrates the potential to identify building materials in a non-destructive, non-contact manner to aid in in-situ facade material labelling.

A Deep Learning Method for Mapping Glacial Lakes from the Combined Use of Synthetic-Aperture Radar and Optical Satellite Images

Glacial lakes (GLs), a vital link between the hydrosphere and the cryosphere, participate in the local hydrological process, and their interannual dynamic evolution is an objective reflection and an indicator of regional climate change. The complex terrain and climatic conditions in mountainous areas where GLs are located make it difficult to employ conventional remote sensing observation means to obtain stable, accurate, and comprehensive observation data. In view of this situation, this study presents an algorithm with a high generalization ability established by optimizing and improving a deep learning (DL) semantic segmentation network model for extracting GL contours from combined synthetic-aperture radar (SAR) amplitude and multispectral imagery data. The aim is to use the high penetrability and all-weather advantages of SAR to reduce the effects of cloud cover as well as to integrate the multiscale and detail-oriented advantages of multispectral data to facilitate accurate, quantitative extraction of GL contours. The accuracy and reliability of the model and algorithm were examined by employing them to extract the contours of GLs in a large region of south-eastern Tibet from Landsat 8 optical remote sensing images and Sentinel-1A amplitude images. In this study, the contours of a total 8262 GLs in south-eastern Tibet were extracted. These GLs were distributed predominantly at altitudes of 4000–5500 m. Only 17.4% of these GLs were greater than 0.1 km2 in size, while a large number of small GLs made up the majority. Through analysis and validation, the proposed method was found highly capable of distinguishing rivers and lakes and able to effectively reduce the misidentification and extraction of rivers. With the DL model based on combined optical and SAR images, the intersection-over-union (IoU) score increased by 0.0212 (to 0.6207) on the validation set and by 0.038 (to 0.6397) on the prediction set. These validation data sufficiently demonstrate the efficacy of the model and algorithm. The technical means employed in this study as well as the results and data obtained can provide a reference for research and application expansion in related fields.

Drone data reveal heterogeneity in tundra greenness and phenology not captured by satellites

Data across scales are required to monitor ecosystem responses to rapid warming in the Arctic and to interpret tundra greening trends. Here, we tested the correspondence among satellite- and drone-derived seasonal change in tundra greenness to identify optimal spatial scales for vegetation monitoring on Qikiqtaruk—Herschel Island in the Yukon Territory, Canada. We combined time-series of the Normalised Difference Vegetation Index (NDVI) from multispectral drone imagery and satellite data (Sentinel-2, Landsat 8 and MODIS) with ground-based observations for two growing seasons (2016 and 2017). We found high cross-season correspondence in plot mean greenness (drone-satellite Spearman’s ρ 0.67–0.87) and pixel-by-pixel greenness (drone-satellite R 2 0.58–0.69) for eight one-hectare plots, with drones capturing lower NDVI values relative to the satellites. We identified a plateau in the spatial variation of tundra greenness at distances of around half a metre in the plots, suggesting that these grain sizes are optimal for monitoring such variation in the two most common vegetation types on the island. We further observed a notable loss of seasonal variation in the spatial heterogeneity of landscape greenness (46.2%–63.9%) when aggregating from ultra-fine-grain drone pixels (approx. 0.05 m) to the size of medium-grain satellite pixels (10–30 m). Finally, seasonal changes in drone-derived greenness were highly correlated with measurements of leaf-growth in the ground-validation plots (mean Spearman’s ρ 0.70). These findings indicate that multispectral drone measurements can capture temporal plant growth dynamics across tundra landscapes. Overall, our results demonstrate that novel technologies such as drone platforms and compact multispectral sensors allow us to study ecological systems at previously inaccessible scales and fill gaps in our understanding of tundra ecosystem processes. Capturing fine-scale variation across tundra landscapes will improve predictions of the ecological impacts and climate feedbacks of environmental change in the Arctic.

Beyond Measurement: Extracting Vegetation Height from High Resolution Imagery with Deep Learning

Measuring and monitoring the height of vegetation provides important insights into forest age and habitat quality. These are essential for the accuracy of applications that are highly reliant on up-to-date and accurate vegetation data. Current vegetation sensing practices involve ground survey, photogrammetry, synthetic aperture radar (SAR), and airborne light detection and ranging sensors (LiDAR). While these methods provide high resolution and accuracy, their hardware and collection effort prohibits highly recurrent and widespread collection. In response to the limitations of current methods, we designed Y-NET, a novel deep learning model to generate high resolution models of vegetation from highly recurrent multispectral aerial imagery and elevation data. Y-NET’s architecture uses convolutional layers to learn correlations between different input features and vegetation height, generating an accurate vegetation surface model (VSM) at 1×1 m resolution. We evaluated Y-NET on 235 km2 of the East San Francisco Bay Area and find that Y-NET achieves low error from LiDAR when tested on new locations. Y-NET also achieves an R2 of 0.83 and can effectively model complex vegetation through side-by-side visual comparisons. Furthermore, we show that Y-NET is able to identify instances of vegetation growth and mitigation by comparing aerial imagery and LiDAR collected at different times.

An updated survey on the use of geospatial technologies in New Zealand’s plantation forestry sector

Background: Geospatial technologies have developed rapidly in recent decades and can provide detailed, accurate data to support forest management. Knowledge of the uptake of geospatial technologies, as well as barriers to adoption, in New Zealand’s plantation forest management sector is limited and would be beneficial to the industry. This study provides an update to the 2013 benchmark study by Morgenroth and Visser. Methods: An online survey was sent to 29 companies that own or manage plantation forests in New Zealand. The survey was split into seven sections, composed of multiple-choice and open-ended questions, on the topics of: demographic information, data portals and datasets, global navigation satellite system (GNSS) receivers, and four remote-sensing technologies. These included aerial imagery, multispectral imagery, hyperspectral imagery, and light detection and ranging (LiDAR). Each section included questions relating to the acquisition, application and products created from each remote-sensing technology. Questions were also included that related to the barriers preventing the uptake of technologies. To determine the progression in the uptake of these technologies the results were compared to Morgenroth and Visser's study conducted five years' earlier. Results: Twenty-three companies responded to the survey and together, those companies managed approximately 1,172,000 ha (or 69% of New Zealand’s 1.706 million ha plantation forest estate (NZFOA, 2018)). The size of the estates managed by individual companies ranged from 1,000 ha to 177,000 ha (quartile 1 = 19,000 ha, median = 33,000 ha, quartile 3 = 63,150 ha). All companies used GNSS receivers and acquired three-band, Red-Green-Blue, aerial imagery. Multispectral imagery, hyperspectral imagery and LiDAR data were acquired by 48%, 9% and 70% of companies, respectively. Common applications for the products derived from these technologies were forest mapping and description, harvest planning, and cutover mapping. The main barrier preventing companies from acquiring most remotely-sensed data was the lack of staff knowledge and training, though cost was the main barrier to LiDAR acquisition. The uptake of all remote-sensing technologies has increased since 2013. LiDAR had the largest progression in uptake, increasing from 17% to 70%. There has also been a change in the way companies acquired the data. Many of the companies used unpiloted aerial vehicles (UAV) to acquire aerial and multispectral imagery in 2018, while in 2013 no companies were using UAVs. ESRI ArcGIS continues to be the dominant geographic information system used by New Zealand’s forest management companies (91%), though 22% of companies now use free GIS software, like QGIS or GRASS. The use of specialised software (e.g. FUSION, LAStools) for LiDAR or photogrammetric point cloud analysis increased since 2013, but most forestry companies who are processing .las files into various products (e.g. digital terrain model) are using ArcGIS. Conclusions: This study showed that there had been a progression in the uptake of geospatial technologies in the New Zealand plantation forest management sector. However, there are still barriers preventing the full utilisation of these technologies. The results suggest that the industry could benefit from investing in more training relating to geospatial technologies.

Shrub Fractional Cover Estimation and Mapping of San Clemente Island Shrubland Based on Airborne Multispectral Imagery and Lidar Data

The purpose of this study is to map shrub distributions and estimate shrub cover fractions based on the classification of high-spatial-resolution aerial orthoimagery and light detection and ranging (LiDAR) data for portions of the highly disturbed coastal sage scrub landscapes of San Clemente Island, California. We utilized nine multi-temporal aerial orthoimage sets for the 2010 to 2018 period to map shrub cover. Pixel-based and object-based image analysis (OBIA) approaches to image classification of growth forms were tested. Shrub fractional cover was estimated for 10, 20 and 40 m grid sizes and assessed for accuracy. The most accurate estimates of shrub cover were generated with the OBIA method with both multispectral brightness values and canopy height estimates from a normalized digital surface model (nDSM). Fractional cover products derived from 2015 and 2017 orthoimagery with nDSM data incorporated yielded the highest accuracies. Major factors that influenced the accuracy of shrub maps and fractional cover estimates include the time of year and spatial resolution of the imagery, the type of classifier, feature inputs to the classifier, and the grid size used for fractional cover estimation. While tracking actual changes in shrub cover over time was not the purpose, this study illustrates the importance of consistent mapping approaches and high-quality inputs, including very-high-spatial-resolution imagery and an nDSM.

Integrating spectral and non-spectral data to improve urban settlement mapping in a large Latin-American city

ABSTRACT Information on urban settlements is crucial for sustainability planning and management. While remote sensing has been used to derive such information, its applicability can be compromised due to the complexity in the urban environment. In this study, we developed a remote sensing method to map land cover types in a large Latin-American city, which is well known for its mushrooming unplanned and informal settlements. After carefully considering the landscape complexity there, we designed a data fusion method combining multispectral imagery and non-spectral data for urban and land mapping. Specifically, we acquired a cloud-free Landsat-8 image and two non-spectral datasets, i.e., digital elevation models and road networks. Then, we implemented a set of experiments with different inputs to evaluate their merits in thematic mapping through a supervised protocol. We found that the map generated with the multispectral data alone had an overall accuracy of 73.3% but combining multispectral imagery and non-spectral data yielded a land cover map with 90.7% overall accuracy. Interestingly, the thermal infrared information helped substantially improve both the overall and categorical accuracies, particularly for the two urban classes. The two types of non-spectral data were critical in resolving several spectrally confused categories, thus considerably increasing the mapping accuracy. However, the panchromatic band with higher spatial resolution and its derived textural measurement only generated a marginal accuracy improvement. The novelties of our work are with the successful separation between the two major types of urban settlements in a complex environment using a carefully designed data fusion approach and the insight into the relative merits of the thermal infrared information and non-spectral data in helping resolve the issue of class ambiguity. These findings should be valuable in deriving accurate urban settlement information which can further advance the research on socio-ecological dynamics and urban sustainability.

ESTIMATION OF SOIL BULK DENSITY AND CARBON USING MULTI-SOURCE REMOTELY SENSED DATA

Abstract. Bulk density and soil carbon models were fitted for soil samples collected during field campaigns in 2018 and 2019 for the Kapuskasing region of the District of Cochrane in Ontario, Canada. Prediction maps for bulk density and soil carbon were generated for the 0–15 cm depth mineral soil layer. The application of multi-source remotely sensed data as environmental covariates for model predictors was implemented. Environmental covariates were obtained from multispectral satellite imagery, LiDAR (light detection and ranging) retrievals and airborne geomagnetic surveys, as well from a digital elevation model (DEM) for topographic covariates. Two covariates derived from LiDAR, canopy height model (CHM) and gap fraction, were of high variable importance when fitting models for average bulk density; gap fraction had the highest to second highest variable importance for average bulk density when considered among a full set of 76, or reduced sets of 12 or 5 separate predictors respectively. Environmental covariates corresponding to vegetation cover, specifically reflectance from multispectral imagery or LiDAR data, had the highest variable importance when compared with other categories of soil formation factors. Random forest (RF) models were generated, with RF models based upon just 12 predictors obtaining reasonable results with coefficients of determinations (R2) greater than 0.7 for the standard derivation of bulk density, standard deviation of total carbon and average total carbon for the 0–15 cm depth layer.

Combining Satellite Multispectral Imagery and Topographic Data for the Detection and Mapping of Fluvial Avulsion Processes in Lowland Areas

Fluvial avulsion is an important process in the dynamics of the riverscapes and plays a key role in the drainage network evolution in lowland areas, also influencing past and present social processes and economic activities. Crevasse splays represent significant geomorphological features for understanding the fluvial morphodynamics in lowland areas dominated by avulsion processes. Within wide floodplains characterized by very low elevation ranges, the detection and accurate mapping of crevasse splay morphology and features, such as crevasse channels, levees, and deposit, can be very challenging considering floodplain extension, anthropic impact on the natural channels network, logistic difficulties, and in some cases, climate conditions that prevent field work. This research aims at improving the detection and mapping of crevasse splays in lowland areas through the combination of different remote sensing techniques based on optical multispectral imagery and topographic data derived from satellite earth observation missions. The Lower Mesopotamia Plain (LMP) offers a unique opportunity to study the avulsion processes because it presents numerous examples of crevasse splays, characterized by different sizes and states of activity. Furthermore, in this area, a strong correlation exists between the formation and development of crevasse splays and the expansion of agriculture and early societies since the Early Holocene. Different supervised classification (SC) methods of Landsat 8 satellite images have been tested together with topographic analysis of the microrelief, carried out based on two different 1-arcsec DEMs (AW3D30 and GDEM2). The results of this study demonstrate that the combination of multispectral imagery analysis and topographic analysis of the microrelief is useful for discerning different crevasse elements, distinguishing between active and relict landforms. The methodological approach proved helpful for improving the mapping of erosional and depositional landforms generated by the avulsion process and, in the study area, provided the best results for the active landforms.

X-ModalNet: A semi-supervised deep cross-modal network for classification of remote sensing data

This paper addresses the problem of semi-supervised transfer learning with limited cross-modality data in remote sensing. A large amount of multi-modal earth observation images, such as multispectral imagery (MSI) or synthetic aperture radar (SAR) data, are openly available on a global scale, enabling parsing global urban scenes through remote sensing imagery. However, their ability in identifying materials (pixel-wise classification) remains limited, due to the noisy collection environment and poor discriminative information as well as limited number of well-annotated training images. To this end, we propose a novel cross-modal deep-learning framework, called X-ModalNet, with three well-designed modules: self-adversarial module, interactive learning module, and label propagation module, by learning to transfer more discriminative information from a small-scale hyperspectral image (HSI) into the classification task using a large-scale MSI or SAR data. Significantly, X-ModalNet generalizes well, owing to propagating labels on an updatable graph constructed by high-level features on the top of the network, yielding semi-supervised cross-modality learning. We evaluate X-ModalNet on two multi-modal remote sensing datasets (HSI-MSI and HSI-SAR) and achieve a significant improvement in comparison with several state-of-the-art methods.

Evaluation of Features Derived from High-Resolution Multispectral Imagery and LiDAR Data for Object-Based Support Vector Machine Classification of Tree Species

Remote sensing can play a key role in understanding the make-up of urban forests. This study analyzes how high-resolution Geoeye-1 multispectral imagery and LiDAR point clouds allow for improved classification of urban tree species using object-based and support vector machine classification (SVM). Five common urban trees are classified: Acer platanoides; Acer platanoides ‘Schwedleri’; Picea pungens; Gleditsia triacanthos; and Tilia cordata. Numerous features are used for classification: index derived from imagery reflectance; texture of imagery; LiDAR height and intensity; and a LiDAR-generated normalized digital surface model. Classification is performed to evaluate the contribution of individual features, groups of features, and the combination of features from both imagery and LiDAR data. Classification results in an overall accuracy of 85.08% when features from both data sources are combined, compared with 77.73% when using only LiDAR features, and 71.85% when using only imagery features.

Fusion of multispectral imagery and LiDAR data for roofing materials and roofing surface conditions assessment

ABSTRACT Assessment of rooftop rainwater harvesting (RRWH) quality and suitability requires detail and reliable information on roofs. Characterization of roof surface conditions affects the quality of harvested rainwater. Nevertheless, the implementation of the system requires improvement in terms of the roof detection techniques to ensure the roof of the building is selected appropriately. Thus, the classification techniques need to be optimized to detect roof materials and roof surface conditions (new or old) with high accuracy. This study aimed to produce high precision detailed roof materials and roof surface conditions map with using high-resolution remote sensing imagery, WorldView-3 (WV3) and light detection and ranging (LiDAR) data. Three different fusion methods; layer stacking (LS), Gram-Schmidt (GS) and principal components spectral sharpening (PCSS) were explored and their performances were compared to improve the spatial and spectral richness of the image. Subsequently, the roof materials and roof surface conditions classes which include old concrete, new concrete, old metal, new metal, old asbestos and new asbestos had been discriminated by employing support vector machine (SVM) and the rule-based technique known as a decision tree (DT). Generally, generated rule-sets present a higher overall accuracy with 87%, 72% and 66% for LS, GS and PCSS, respectively. For SVM classifier, the maximum accuracy recorded for LS, PCSS and GS were 70%, 63% and 43% respectively. Therefore, rule-based classification via LS fusion technique was utilized to identify suitable rooftops for the development of harvested rainwater system in the urban area. Findings indicate that the degradation status of a roof in heterogenous urban environments could be determined from satellite observation and the quality of roof-based harvested rainwater affected by roofing materials and roofing surface conditions can be analysed effectively.

Influence of selected habitat and stand factors on bark beetle Ips typographus (L.) outbreak in the Białowieża Forest

The European spruce bark beetle (Ips typographus L.) is one of the most critical insect disturbance agents in Europe. In recent years, bark beetles have caused tremendous economic losses, and affected ecosystems over large spatial scales. In this study, we evaluated the influence of selected factors on the bark beetle outbreak in 2015, the year with the most severe drought to have been recorded in the last few decades. The main aim of this study was to develop a new approach of identifying and mapping individual trees infested by bark beetles, for more efficient bark beetle management. The study was conducted on an area occupying 62,000 ha, using airborne laser scanning, multispectral imagery data, and digital forest maps.First, an individual dead tree detection method, based on remote sensing data, was developed, which allowed the quantification of the bark beetle outbreak at an individual tree level, with accurate information on the number and location of killed spruces. Then, topography, habitat, and single tree-based variables were used to examine their influence on tree mortality. Finally, a spatial hot-spot analysis of the outbreak, throughout the 62,000 ha of the Polish part of the Białowieża Forest, was performed.For the first time, we mapped the status of the spruce bark beetle outbreak across the Białowieża Forest, using the method developed in this study. 283,166 dead Norway spruce trees were detected in the study area, supporting the fact that the current bark beetle outbreak is the largest in this area since 1945. The number of dead spruces surrounding any given dead tree was the single most important predictor in all models, with a relative contribution of 35–79% depending on the neighbourhood analysed (radius from 0 to 2000 m). Crown closure and stand age, as well as the share of dead spruce and dominant tree species were found to be the most important predictors of the development of bark beetle infestation. The method developed in this study was evaluated and proved to be suitable for the monitoring and management of the ongoing insect outbreak.

Evaluating Empirical Regression, Machine Learning, and Radiative Transfer Modelling for Estimating Vegetation Chlorophyll Content Using Bi-Seasonal Hyperspectral Images

Different types of methods have been developed to retrieve vegetation attributes from remote sensing data, including conventional empirical regressions (i.e., linear regression (LR)), advanced empirical regressions (e.g., multivariable linear regression (MLR), partial least square regression (PLSR)), machine learning (e.g., random forest regression (RFR), decision tree regression (DTR)), and radiative transfer modelling (RTM, e.g., PROSAIL). Given that each algorithm has its own strengths and weaknesses, it is essential to compare them and evaluate their effectiveness. Previous studies have mainly used single-date multispectral imagery or ground-based hyperspectral reflectance data for evaluating the models, while multi-seasonal hyperspectral images have been rarely used. Extensive spectral and spatial information in hyperspectral images, as well as temporal variations of landscapes, potentially influence the model performance. In this research, LR, PLSR, RFR, and PROSAIL, representing different types of methods, were evaluated for estimating vegetation chlorophyll content from bi-seasonal hyperspectral images (i.e., a middle- and a late-growing season image, respectively). Results show that the PLSR and RFR generally performed better than LR and PROSAIL. RFR achieved the highest accuracy for both images. This research provides insights on the effectiveness of different models for estimating vegetation chlorophyll content using hyperspectral images, aiming to support future vegetation monitoring research.

Evaluation of Grass Quality under Different Soil Management Scenarios Using Remote Sensing Techniques

Hyperspectral and multispectral imagery have been demonstrated to have a considerable potential for near real-time monitoring and mapping of grass quality indicators. The objective of this study was to evaluate the efficiency of remote sensing techniques for quantification of aboveground grass biomass (BM) and crude protein (CP) in a temperate European climate such as Ireland. The experiment was conducted on 64 plots and 53 paddocks with varying quantities of nitrogen applied. Hyperspectral imagery (HSI) and multispectral imagery (MSI) were analyzed to develop the prediction models. The MSI data used in this study were captured using an unmanned aircraft vehicle (UAV) and the satellite Sentinel-2, while the HSI data were obtained using a handheld hyperspectral camera. The prediction models were developed using partial least squares regression (PLSR) and stepwise multi-linear regression (MLR). Eventually, the spatial distribution of grass biomass over plots and paddocks was mapped to assess the within-field variability of grass quality metrics. An excellent accuracy was achieved for the prediction of BM and CP using HSI (RPD > 2.5 and R2 > 0.8), and a good accuracy was obtained via MSI-UAV (2 < RPD < 2.5 and R2 > 0.7) for the grass quality indicators. The accuracy of the models calculated using MSI-Sentinel-2 was reasonable for BM prediction and insufficient for CP estimation. The red-edge range of the wavelengths showed the maximum impact on the predictability of grass BM, and the NIR range had the greatest influence on the estimation of grass CP. Both the PLSR and MLR techniques were found to be sufficiently robust for spectral modelling of aboveground BM and CP. The PLSR yielded a slightly better model than MLR. This study suggested that remote sensing techniques can be used as a rapid and reliable approach for near real-time quantitative assessment of fresh grass quality under a temperate European climate.

INTEGRATION OF A LOW-COST MULTISENSORY UAV SYSTEM FOR FOREST APPLICATION

Abstract. To integrate the multi-spectral imagery and laser scanning data for forest management, a low-cost multisensory UAV system, named Kylin Cloud, is introduced in this paper. The Kylin Cloud is composed several low-cost sensors (i.e., GNSS receiver, IMU, global shutter camera, multispectral camera, and laser scanner), providing the fusion of the imagery and laser scanning data for reliable forest inventory. Experiments were undertaken in a forest park in Wuhan. Results showed that the registration error of the multispectral Digital Orthophoto Map (DOM) and laser scanning data is about one pixel, demonstrating a high potential of the proposed low-cost system.

A comparison of multiple methods for mapping local-scale mesquite tree aboveground biomass with remotely sensed data

A local-scale mesquite tree (Prosopis glandulosa Torr.) aboveground biomass map contribute to our understanding of the spatial distribution of woody plant aboveground biomass, and carbon stocks and fluxes in rangeland ecosystems. The objective of the study was examining a methodological approach to use airborne lidar data and multispectral imagery to create very high spatial resolution local-scale mesquite tree aboveground biomass maps by comparing three statistical methods and identifying significant prediction variables. The three statistical methods were the stepwise regression, the least absolute shrinkage and selection operator (LASSO), and the random forests. These methods were applied to establish the mesquite tree aboveground biomass equations and model from the in-situ mesquite tree aboveground biomass with the lidar metrics and multispectral data. The results showed the stepwise regression and LASSO had limited adj-R2 and MSE. However, the random forests method with combined multispectral imagery and lidar data presented acceptable MSE and R2 (1.08 Mg ha−1 and 0.37). In summary, the random forests method with combined multispectral imagery and lidar data offered the most reliable and reasonable combination to generate a very high spatial resolution local-scale mesquite tree aboveground biomass map.