UAV-based Hyperspectral Imaging Research Articles

Remote Prediction of Soybean Yield Using UAV-Based Hyperspectral Imaging and Machine Learning Models

Early soybean yield estimation has become a fundamental tool for market policy and food security. Considering a heterogeneous crop, this study investigates the spatial and spectral variability in soybean canopy reflectance to achieve grain yield estimation. Besides allowing crop mapping, remote sensing data also provide spectral evidence that can be used as a priori knowledge to guide sample collection for prediction models. In this context, this study proposes a sampling design method that distributes sample plots based on the spatial and spectral variability in vegetation spectral indices observed in the field. Random forest (RF) and multiple linear regression (MLR) approaches were applied to a set of spectral bands and six vegetation indices to assess their contributions to the soybean yield estimates. Experiments were conducted with a hyperspectral sensor of 25 contiguous spectral bands, ranging from 500 to 900 nm, carried by an unmanned aerial vehicle (UAV) to collect images during the R5 soybean growth stage. The tests showed that spectral indices specially designed from some bands could be adopted instead of using multiple bands with MLR. However, the best result was obtained with RF using spectral bands and the height attribute extracted from the photogrammetric height model. In this case, Pearson’s correlation coefficient was 0.91. The difference between the grain yield productivity estimated with the RF model and the weight collected at harvest was 1.5%, indicating high accuracy for yield prediction.

Development and Application of Unmanned Aerial High-Resolution Convex Grating Dispersion Hyperspectral Imager.

This study presents the design and development of a high-resolution convex grating dispersion hyperspectral imaging system tailored for unmanned aerial vehicle (UAV) remote sensing applications. The system operates within a spectral range of 400 to 1000 nm, encompassing over 150 channels, and achieves an average spectral resolution of less than 4 nm. It features a field of view of 30°, a focal length of 20 mm, a compact volume of only 200 mm × 167 mm × 78 mm, and a total weight of less than 1.5 kg. Based on the design specifications, the system was meticulously adjusted, calibrated, and tested. Additionally, custom software for the hyperspectral system was independently developed to facilitate functions such as control parameter adjustments, real-time display, and data preprocessing of the hyperspectral camera. Subsequently, the prototype was integrated onto a drone for remote sensing observations of Spartina alterniflora at Yangkou Beach in Shouguang City, Shandong Province. Various algorithms were employed for data classification and comparison, with support vector machine (SVM) and neural network algorithms demonstrating superior classification accuracy. The experimental results indicate that the UAV-based hyperspectral imaging system exhibits high imaging quality, minimal distortion, excellent resolution, an expansive camera field of view, a broad detection range, high experimental efficiency, and remarkable capabilities for remote sensing detection.

Unsupervised Characterization of Water Composition with UAV-Based Hyperspectral Imaging and Generative Topographic Mapping

Unmanned aerial vehicles equipped with hyperspectral imagers have emerged as an essential technology for the characterization of inland water bodies. The high spectral and spatial resolutions of these systems enable the retrieval of a plethora of optically active water quality parameters via band ratio algorithms and machine learning methods. However, fitting and validating these models requires access to sufficient quantities of in situ reference data which are time-consuming and expensive to obtain. In this study, we demonstrate how Generative Topographic Mapping (GTM), a probabilistic realization of the self-organizing map, can be used to visualize high-dimensional hyperspectral imagery and extract spectral signatures corresponding to unique endmembers present in the water. Using data collected across a North Texas pond, we first apply GTM to visualize the distribution of captured reflectance spectra, revealing the small-scale spatial variability of the water composition. Next, we demonstrate how the nodes of the fitted GTM can be interpreted as unique spectral endmembers. Using extracted endmembers together with the normalized spectral similarity score, we are able to efficiently map the abundance of nearshore algae, as well as the evolution of a rhodamine tracer dye used to simulate water contamination by a localized source.

Morphological Dune Mapping in Shallow Alluvial Stream Using UAV-based Hyperspectral Images

Characterizing morphological features in shallow streams such as dunes and ripples is vital to studies on fluvial geomorphology and in-stream habitat assessments of stream ecology. The paper aimed to examine the feasibility of a conventional hyperspectral method called linear optimal band ratio analysis for capturing the detailed morphologies in shallow small streams, allowing the identification of ripples and dunes. The present study involved a dedicated field experiment at the Gam stream, which is a tributary of the Nakdong River, South Korea. An unmanned aerial vehicle based hyperspectral image was obtained with a spatial resolution of <10 cm and developed an optimal depth-band ratio rating by densely scattered in situ bathymetry measurements with a portable Real Time Kinematic Global Positioning System. The derived hyperspectral bathymetric map with a 7 cm spatial resolution successfully captured the detailed bed morphology, where dunes with sizes of 1.5 m were clearly identifiable. The correlation with depth measured by RTK-GPS was found to be 0.956, with Root Mean Square Error of 0.033 meter. The research confirmed that the conventional linear OBRA used for low-altitude UAV-based hyperspectral images can capture morphological features in shallow streams with a high spatial resolution.

A 3D-2DCNN-CA approach for enhanced classification of hickory tree species using UAV-based hyperspectral imaging

Hickory trees possess substantial economic, nutritional, and ecological value, and they play a pivotal role in both human societies and natural ecosystems. To maximize benefits, many bases have adopted a composite planting strategy. Effective and timely monitoring of hickory forest species information is crucial for precise management and conservation. In this study, a low-altitude unmanned aerial vehicle (UAV) was employed to capture RGB and hyperspectral data from the canopy of two hickory bases. A hybrid convolutional neural network structure was then utilized to classify different tree species and congeneric hickory species. By introducing a channel attention module to refine the features of hyperspectral images, classification stability was enhanced. The experimental results demonstrate that hyperspectral images yield superior classification performance in hickory species identification compared to RGB images, particularly in classifying highly homogeneous tree species. The 3D-2DCNN-CA proposed in this paper demonstrated the best performance in classifying hickory species compared to other classification methods. In the classification of different tree species, the accuracy for a single hickory reached 99.38%, and in the classification of the same hickory species, it achieved an accuracy of 93.97%. Furthermore, this method also achieved significant classification results at the single-tree scale. These results indicate that the method can realize fine-scale monitoring of hickory forests and provide substantial support for forest land management and expert guidance on planting distribution.

Evaluation of Growth Recovery Grade in Lodging Maize via UAV-Based Hyperspectral Images

Rapid and nondestructive monitoring of the temporal dynamic changes of agronomic traits of lodging maize is crucial for evaluating the growth recovery status. The purpose of this study is to assess the time-series changes in maize growth recovery after lodging using unmanned aerial vehicle (UAV) hyperspectral technology. Based on the Entropy method, canopy height (CH) and canopy coverage (CC) were used to represent the canopy structure index (CSI), while leaf chlorophyll content (LCC) and plant water content (PWC) were used to represent the physiological activity index (PAI). Based on the theory of normal (skewed) distribution, the growth recovery grade (GRG) of lodging maize was divided based on the estimated CSI and PAI values. The main results were as follows: (a) With the advance of days after lodging (DAL), CH was decreased after increasing, while other agronomic traits exhibited a downward trend. (b) The R 2 values for the CH, CC, LCC, and PWC estimation model were 0.75, 0.69, 0.54, and 0.49, respectively, while the MAPE values were 14.03%, 8.84%, 16.62%, and 6.22%, respectively, in the testing set. (c) The growth recovery of lodging maize was classified using the threshold based on estimated CSI and PAI, achieving an overall accuracy of 77.68%. Therefore, the method for evaluating maize growth recovery after lodging proved effective in monitoring lodging damage. This study provided a reference for the efficient and nondestructive monitoring of growth recovery in lodging maize using UAV-based hyperspectral images.

Comparison of different machine learning algorithms for predicting maize grain yield using UAV-based hyperspectral images

Comparison of different machine learning algorithms for predicting maize grain yield using UAV-based hyperspectral images

Monitoring of chlorophyll-a and suspended sediment concentrations in optically complex inland rivers using multisource remote sensing measurements

In recent decades, phytoplankton proliferation and sediment input to rivers (especially urban rivers) have become more dramatic under the compound pressure of climate change and human activities. Given the generally narrow width of rivers and current high spatial resolution satellites, which are limited by band settings, bandwidth, and the signal-to-noise ratio, UAVs with their exceptional spatiotemporal resolution can be used as a useful tool for river environmental monitoring and inversion uncertainty assessment. In this study, UAV-based hyperspectral (X20P) and multispectral (P4M) images, along with Sentinel-2 MultiSpectral Instrument (MSI), Landsat-8 Operational Land Imager (OLI) and Landsat-9 OLI2 data, were used to assess the uncertainty in retrieving chlorophyll-a (Chla) and suspended sediment (SS) concentrations in rivers. Chla and SS models based on UAV and satellite data were constructed using stepwise multiple regression and typical Chla and SS retrieval algorithms, respectively, and the performance of the models was the focus of our research. The results demonstrated that in the Chla concentration inversion, each sensor performed as follows: X20P > P4M > Landsat9 OLI2 > Sentinel-2 MSI > Landsat8 OLI, and the performance in the SS concentration inversion was as follows: X20P > Sentinel-2 MSI > P4M > Landsat9 OLI2 > Landsat8 OLI. In addition, the uncertainty of high spatial resolution satellite retrievals was analyzed with the assistance of the UAV-based model. Results showed that narrow bandwidths and finely tuned band settings are more essential for the Chla inversion. The typical Chla retrieval algorithm, NDCI, is only effective in certain bands (band 1 from 684 to 724 nm and band 2 from 660 to 680 nm). It is also noted that Landsat8 and Landsat9 lack some key band settings (e.g., the red-edge band of 700–710 nm), severely limiting practical application in relation to Chla. However, specific variances in different sensor bands have a relatively small impact on SS inversion, for example, the correlation between SS and the R/B (a typical SS retrieval algorithm) constructed by each sensor ranged from 0.68 to 0.77. Chla monitoring, on the other hand, necessitates a higher spatial resolution than SS monitoring. The accuracy decreased markedly when UAV images were resampled to 10 m and 30 m spatial resolution. However, it is not as crucial for the SS inversion, images with the original spatial resolution (RMSE<30cm = 6.28 mg/L) were resampled to 10 m resolution (RMSE10m = 5.85 mg/L) and 30 m resolution (RMSE30m = 4.08 mg/L) while using P4M for SS inversion, and the accuracy increased. Our results demonstrated and highlighted various options for future monitoring of Chla and SS, while exploiting the synergy between UAVs and satellites to achieve more precise observations at greater spatial and temporal scales, which will benefit aquatic environment management and protection.

A spectral decomposition method for estimating the leaf nitrogen status of maize by UAV-based hyperspectral imaging

Leaf nitrogen status plays a crucial role in characterizing maize nutrient activity, which ultimately affects both the photosynthetic efficiency and yield formation of maize. For this reason, hyperspectral imaging technology based on unmanned aerial vehicle (UAV) has emerged as a popular tool for estimating crop phenotypic traits. Canopy structure and leaf nutrition together determine the crop canopy spectrum. Efficient separation of the spectral information sensitive to the leaf nitrogen status from the canopy spectrum is considered of utmost importance for improving the estimation accuracy of maize leaf nitrogen status. Along these lines, the main goal of this work was to develop a canopy spectral decomposition method to reduce the interference of other traits on leaf nitrogen concentration (LNC) and leaf nitrogen density (LND) estimation using UAV-based hyperspectral images. First, the weights of the leaf nitrogen status, aboveground biomass (AGB), and leaf area index (LAI) contributing to the canopy spectrum were calculated by using the entropy weight method. Then, the sensitive bands of LNC and LND before and after spectral decomposition were selected by implementing the successive projections algorithm (SPA) and competitive adaptive reweighted sampling (CARS) algorithm. Finally, the reflectance of all bands and sensitive bands was compared to estimate maize LNC and LND using partial least squares regression (PLSR), Gaussian process regression (GPR), support vector regression (SVR), and random forest regression (RFR). These models were systematically tested using independent samples. From the acquired results, it was demonstrated that the correlation coefficient (r) between LNC and each band increased after spectral decomposition compared to the correlation before spectral decomposition. The r between the band reflectance in the near infrared region and LNC or LND after spectral decomposition was significantly higher than that before spectral decomposition. In addition, the sensitive bands of maize leaf nitrogen status after spectral decomposition were around 470 nm, 538 nm, 638 nm, 682 nm, 710 nm, 734 nm, and 830 nm. Regardless of using the reflectance of all bands or sensitive bands, the four estimation models of LNC and LND after spectral decomposition performed better than those before spectral decomposition. The estimation models of LNC and LND constructed by CARS-SVR can successfully reproduce the estimation accuracies of the models constructed by using all-bands-SVR, with R2 on the testing set of 0.68. Theresultshighlight that spectral decomposition is an effective method to significantly improve the estimation accuracy of maize leaf nitrogen status using UAV hyperspectral images, thus effectively reducing the interference of canopy structure traits (AGB and LAI).

UAV-Based Hyperspectral Imaging for River Algae Pigment Estimation

Harmful and nuisance algal blooms are becoming a greater concern to public health, riverine ecosystems, and recreational uses of inland waterways. Algal bloom proliferation has increased in the Upper Clark Fork River due to a combination of warming water temperatures, naturally high phosphorus levels, and an influx of nitrogen from various sources. To improve understanding of bloom dynamics and how they affect water quality, often measured as algal biomass measured through pigment standing crops, a UAV-based hyperspectral imaging system was deployed to monitor several locations along the Upper Clark Fork River in western Montana. Image data were collected across the spectral range of 400–1000 nm with 2.1 nm spectral resolution during two field sampling campaigns in 2021. Included are methods to estimate chl a and phycocyanin standing crops using regression analysis of salient wavelength bands, before and after separating the pigments according to their growth form. Estimates of chl a and phycocyanin standing crops generated through a linear regression analysis are compared to in situ data, resulting in a maximum R2 of 0.96 for estimating fila/epip chl-a and 0.94 when estimating epiphytic phycocyanin. Estimates of pigment standing crops from total abundance, epiphytic, and the sum of filamentous and epiphytic sources are also included, resulting in a promising method for remotely estimating algal standing crops. This method addresses the shortcomings of current monitoring techniques, which are limited in spatial and temporal scale, by proposing a method for rapid collection of high-spatial-resolution pigment abundance estimates.

Utilizing UAV-based hyperspectral imaging to detect surficial explosive ordnance

Across postconflict regions of the world, explosive ordnance (EO), which includes remnant antipersonnel land mines, antivehicle/tank mines, unexploded cluster munitions, improvised explosive devices, and explosive remnants of war (ERW) such as unexploded ordnance and abandoned explosive ordnance, remains a critical humanitarian concern. Clearance and land release efforts anchored on manual geophysical detection and mechanical probing methods remain painstakingly slow, expensive, and dangerous to operators. As a result, postconflict regions impacted by EO contamination significantly lag in social and economic development. Developing, calibrating, and field testing more efficient detection methods for surficial EO is a crucial task. Unpiloted aerial systems featuring advanced remote sensing capabilities are a key technology that may allow the tide to turn in the EO crisis. Specifically, recent advances in hardware design have allowed for effective deployment of small, light, and less power consuming hyperspectral imaging (HSI) systems from small unpiloted aerial vehicles (UAVs). Our proof-of-concept study employs UAV-based HSI to deliver a safer, faster, and more cost-efficient method of surface land mine and ERW detection compared to current ground-based detection methods. Our results indicate that analysis of HSI data sets can produce spectral profiles and derivative data products to distinguish multiple ERW and mine types in a variety of host environments.

A Standardized Procedure to Build a Spectral Library for Hazardous Chemicals Mixed in River Flow Using Hyperspectral Image

The occurrence of natural disasters as a consequence of accidental hazardous chemical spills remains a concern. The inadequate, or delayed, initial response may fail to mitigate their impact; hence, imminent monitoring of responses in the initial stage is critical. Classical contact-type measurement methods, however, sometimes miss solvent chemicals and invoke risks for operators during field operation. Remote sensing methods are an alternative method as non-contact, spatially distributable, efficient and continuously operatable features. Herein, we tackle challenges posed by the increasingly available UAV-based hyperspect ral images in riverine environments to identify the presence of hazardous chemical solvents in rivers, which are less investigated in the absence of direct measurement strategies. We propose a referable standard procedure for a unique spectral library based on pre-scanning hyperspectral sensors with respect to representative hazardous chemicals registered on the national hazardous chemical list. We utilized the hyperspectral images to identify 18 types of hazardous chemicals injected into the river in an outdoor environment, where a dedicated hyperspectral ground imaging system mounted with a hyperspectral camera was designed and applied. Finally, we tested the efficiency of the library to recognize unknown chemicals, which showed &gt;70% success rate.

Machine Learning-Based Ensemble Band Selection for Early Water Stress Identification in Groundnut Canopy Using UAV-Based Hyperspectral Imaging

This paper presents the early identification of water stress in groundnut canopy using unmanned aerial vehicle (UAV) based hyperspectral imaging (in 385-1020 nm) and machine learning (ML) techniques. An efficient hyperspectral imaging (HSI) data analysis pipeline was presented which includes image quality assessment, denoising, band selection, and classification. A novel ML-based ensemble feature selection (FS) algorithm has been proposed for optimal water stress sensitive waveband selection. The data analysis pipeline and the selected bands were validated on HSI data acquired at two different water stress levels. Wavelengths 515.05, 552.16, 711.92, 724.75, and 931.92 nm were identified as optimal water stress sensitive bands in groundnut canopy, using which we could identify early stress with 96.46% accuracy. The proposed data analysis pipeline and ensemble FS algorithm will benefit crop phenotyping applications such as early abiotic stress detection.

Non-destructive monitoring of amylose content in rice by UAV-based hyperspectral images.

Amylose content (AC) is an important indicator for rice quality grading. The rapid development of unmanned aerial vehicle (UAV) technology provides rich spectral and spatial information on observed objects, making non-destructive monitoring of crop quality possible. To test the potential of UAV-based hyperspectral images in AC estimation, in this study, observations on five rice cultivars were carried out in eastern China (Zhejiang province) for four consecutive years (from 2017 to 2020). The correlations between spectral and textural variables of UAV-based hyperspectral images at different growth stages (booting, heading, filling, and ripening) and AC (%) were analyzed, and the linear regression models based on spectral variables alone, textural variables alone, and combined spectral and textural variables were established. The results showed that the sensitive bands (P< 0.001) to AC were mainly centered in the green (536∽568 nm) and red regions (630∽660nm), with spectral and textural variables at the ripening stage giving the highest negative correlation coefficient of -0.868 and -0.824, respectively. Models based on combined spectral and textural variables give better estimation than those based on spectral or textural variables alone, characterized by less variables and higher accuracy. The best models using spectral or textural variables alone both involved three growth stages (heading, filling, and ripening), with root mean square error (RMSE) of 1.01% and 1.04%, respectively, while the models based on combined spectral and textural variables have RMSE of 1.04% 0.844% with only one (ripening stage) or two (ripening and filling stages) growth stages involved. The combination of spectral and textural variables of UAV-based hyperspectral images is expected to simplify data acquisition and enhance estimation accuracy in remote sensing of rice AC.

Integrating the Textural and Spectral Information of UAV Hyperspectral Images for the Improved Estimation of Rice Aboveground Biomass

The accurate and rapid estimation of the aboveground biomass (AGB) of rice is crucial to food security. Unmanned aerial vehicles (UAVs) mounted with hyperspectral sensors can obtain images of high spectral and spatial resolution in a quick and effective manner. Integrating UAV-based spatial and spectral information has substantial potential for improving crop AGB estimation. Hyperspectral remote-sensing data with more continuous reflectance information on ground objects provide more possibilities for band selection. The use of band selection for the spectral vegetation index (VI) has been discussed in many studies, but few studies have paid attention to the band selection of texture features in rice AGB estimation. In this study, UAV-based hyperspectral images of four rice varieties in five nitrogen treatments (N0, N1, N2, N3, and N4) were obtained. First, multiple spectral bands were used to identify the optimal bands of the spectral vegetation indices, as well as the texture features; next, the vegetation index model (VI model), the vegetation index combined with the corresponding-band textures model (VI+CBT model), and the vegetation index combined with the full-band textures model (VI+FBT model) were established to compare their respective rice AGB estimation abilities. The results showed that the optimal bands of the spectral and textural information for AGB monitoring were inconsistent. The red-edge and near-infrared bands demonstrated a strong correlation with the rice AGB in the spectral dimension, while the green and red bands exhibited a high correlation with the rice AGB in the spatial dimension. The ranking of the monitoring accuracies of the three models, from highest to lowest, was: the VI+FBT model, then the VI+CBT model, and then the VI model. Compared with the VI model, the R2 of the VI+FBT model and the VI+CBT model increased by 1.319% and 9.763%, respectively. The RMSE decreased by 2.070% and 16.718%, respectively, while the rRMSE decreased by 2.166% and 16.606%, respectively. The results indicated that the integration of vegetation indices and textures can significantly improve the accuracy of rice AGB estimation. The full-band textures contained richer information that was highly related to rice AGB. The VI model at the tillering stage presented the greatest sensitivity to the integration of textures, and the models in the N3 treatment (1.5 times the normal nitrogen level) gave the best AGB estimation compared with the other nitrogen treatments. This research proposes a reliable modeling framework for monitoring rice AGB and provides scientific support for rice-field management.

UAV-Borne Hyperspectral Imaging Remote Sensing System Based on Acousto-Optic Tunable Filter for Water Quality Monitoring

Unmanned aerial vehicle (UAV) hyperspectral remote sensing technologies have unique advantages in high-precision quantitative analysis of non-contact water surface source concentration. Improving the accuracy of non-point source detection is a difficult engineering problem. To facilitate water surface remote sensing, imaging, and spectral analysis activities, a UAV-based hyperspectral imaging remote sensing system was designed. Its prototype was built, and laboratory calibration and a joint air–ground water quality monitoring activity were performed. The hyperspectral imaging remote sensing system of UAV comprised a light and small UAV platform, spectral scanning hyperspectral imager, and data acquisition and control unit. The spectral principle of the hyperspectral imager is based on the new high-performance acousto-optic tunable (AOTF) technology. During laboratory calibration, the spectral calibration of the imaging spectrometer and image preprocessing in data acquisition were completed. In the UAV air–ground joint experiment, combined with the typical water bodies of the Yangtze River mainstream, the Three Gorges demonstration area, and the Poyang Lake demonstration area, the hyperspectral data cubes of the corresponding water areas were obtained, and geometric registration was completed. Thus, a large field-of-view mosaic and water radiation calibration were realized. A chlorophyl-a (Chl-a) sensor was used to test the actual water control points, and 11 traditional Chl-a sensitive spectrum selection algorithms were analyzed and compared. A random forest algorithm was used to establish a prediction model of water surface spectral reflectance and water quality parameter concentration. Compared with the back propagation neural network, partial least squares, and PSO-LSSVM algorithms, the accuracy of the RF algorithm in predicting Chl-a was significantly improved. The determination coefficient of the training samples was 0.84; root mean square error, 3.19 μg/L; and mean absolute percentage error, 5.46%. The established Chl-a inversion model was applied to UAV hyperspectral remote sensing images. The predicted Chl-a distribution agreed with the field observation results, indicating that the UAV-borne hyperspectral remote sensing water quality monitoring system based on AOTF is a promising remote sensing imaging spectral analysis tool for water.

Using UAV-based hyperspectral imaging and functional regression to assist in predicting grain yield and related traits in wheat under heat-related stress environments for the purpose of stable yielding genotypes

Quantifying certain physiological traits under heat-stress is crucial for maximizing genetic gain for wheat yield and yield-related components. In-season estimation of different physiological traits related to heat stress tolerance can ensure the finding of germplasm, which could help in making effective genetic gains in yield. However, estimation of those complex traits is time- and labor-intensive. Unmanned aerial vehicle (UAV) based hyperspectral imaging could be a powerful tool to estimate indirectly in-season genetic variation for different complex physiological traits in plant breeding that could improve genetic gains for different important economic traits, like grain yield. This study aims to predict in-season genetic variations for cellular membrane thermostability (CMT), yield and yield related traits based on spectral data collected from UAVs; particularly, in cases where there is a small sample size to collect data from and a large range of features collected per sample. In these cases, traditional methods of yield-prediction modeling become less robust. To handle this, a functional regression approach was employed that addresses limitations of previous techniques to create a model for predicting CMT, grain yield and other traits in wheat under heat stress environmental conditions and when data availability is constrained. The results preliminarily indicate that the overall models of each trait studied presented a good accuracy compared to their data’s standard deviation. The yield prediction model presented an average error of 13.42%, showing the function-on-function algorithm chosen for the model as reliable for small datasets with high dimensionality.

Using UAV-Based Hyperspectral Imagery to Detect Winter Wheat Fusarium Head Blight

Fusarium head blight (FHB) is a major winter wheat disease in China. The accurate and timely detection of wheat FHB is vital to scientific field management. By combining three types of spectral features, namely, spectral bands (SBs), vegetation indices (VIs), and wavelet features (WFs), in this study, we explore the potential of using hyperspectral imagery obtained from an unmanned aerial vehicle (UAV), to detect wheat FHB. First, during the wheat filling period, two UAV-based hyperspectral images were acquired. SBs, VIs, and WFs that were sensitive to wheat FHB were extracted and optimized from the two images. Subsequently, a field-scale wheat FHB detection model was formulated, based on the optimal spectral feature combination of SBs, VIs, and WFs (SBs + VIs + WFs), using a support vector machine. Two commonly used data normalization algorithms were utilized before the construction of the model. The single WFs, and the spectral feature combination of optimal SBs and VIs (SBs + VIs), were respectively used to formulate models for comparison and testing. The results showed that the detection model based on the normalized SBs + VIs + WFs, using min–max normalization algorithm, achieved the highest R2 of 0.88 and the lowest RMSE of 2.68% among the three models. Our results suggest that UAV-based hyperspectral imaging technology is promising for the field-scale detection of wheat FHB. Combining traditional SBs and VIs with WFs can improve the detection accuracy of wheat FHB effectively.

Estimating the maize biomass by crop height and narrowband vegetation indices derived from UAV-based hyperspectral images

Monitoring the aboveground biomass (AGB) of maize is essential for improving site-specific nutrient management and predicting yield to ensure food safety. A low-altitude unmanned aerial vehicle (UAV) was employed to acquire hyperspectral imagery of the maize canopy at three growth stages (V6, R1, R3) to estimate the maize AGB. Five maize nitrogen (N) rate experiments were conducted in Lishu County, Jilin Province, Northeastern China, to create different biomass conditions. Combined with crop height data obtained from the field measurements, 30 narrowband vegetation indices (VIs) were extracted using surface reflectance data from the hyperspectral imagery. Stepwise regression, random forest (RF) regression and XGBoost regression models were used to predict the fresh and dry AGB of the 2019 growth season. The study revealed that (1) crop height explained the most variability (60–70%) in the maize dry and fresh AGB estimation of the V6 growth stage, while different VIs exhibited various importance for AGB estimation at other maize growth stages; (2) XGBoost regression models demonstrated high prediction accuracy in both fresh and dry AGB estimation, compared with stepwise regression and RF models. (3) XGBoost models also presented high prediction accuracy at each single-growth stage and the whole-growth stage, with the highest accuracy for the dry AGB at V6 growth stage (R2 = 0.81, RMSE = 0.27 t/ha). This study demonstrated the capability of UAV-based hyperspectral imagery for estimating maize AGB at the field scale, which can be used to assist precision agriculture.

Combining spectral and textural information in UAV hyperspectral images to estimate rice grain yield

The speedy development of UAV (Unmanned Aerial Vehicle) has provided more data choices for crop yield estimation. In most cases, spectral information derived from multispectral or hyperspectral images is used alone for yield estimation, while variables that contain spatial information such as textural measures were not considered. As UAV can acquire images with spatial resolution at centimeter-level, which contain rich spatial information of observed objects, the test of textural measures in improving yield estimation becomes possible. In this study, UAV-based hyperspectral images of rice in the east of China were acquired in 2017, 2018 and 2019 three consecutive year during rice growing seasons. Eight gray level co-occurrence matrix (GLCM) based texture measures were derived, and all two band combinations of these textural measures as well as spectral reflectance in three types (normalized type, differential type and simple ratio type) were calculated to developing textural and spectral index. Grain yield estimation models were established using spectral indices alone and spectral indices combining textural measures, respectively. The results show that models based on spectral indices combining textural measures all performed better than these using spectral indices alone, with coefficient of determination R2 of the best model greater than 0.8, RMSE (Root Mean Square Error) of 0.421 Mg ha−1 and MAPE (Mean Absolute Percentage Error) of 4.66% for calibration results, and 0.521 Mg ha−1 and 6.63% for validation results. Meanwhile, spectral indices combining textural measures can adjust the problems of overestimation and underestimation in models developed with spectral indices alone. This kind of improvement is especially obvious when grain yield is below 5.0 Mg ha−1 or above 8.0 Mg ha−1. Besides, the estimation accuracy of the model based on spectral index combining textural measure at the booting stage was equivalent to the model based on spectral indices at the booting and heading stages, indicating that textural measures may be helpful in monitoring grain yield at an earlier stage.