UAV-Based Hyperspectral and Ensemble Machine Learning for Predicting Yield in Winter Wheat

Winter wheat is a widely-grown cereal crop worldwide. Using growth-stage information to estimate winter wheat yields in a timely manner is essential for accurate crop management and rapid decision-making in sustainable agriculture, and to increase productivity while reducing environmental impact. UAV remote sensing is widely used in precision agriculture due to its flexibility and increased spatial and spectral resolution. Hyperspectral data are used to model crop traits because of their ability to provide continuous rich spectral information and higher spectral fidelity. In this study, hyperspectral image data of the winter wheat crop canopy at the flowering and grain-filling stages was acquired by a low-altitude unmanned aerial vehicle (UAV), and machine learning was used to predict winter wheat yields. Specifically, a large number of spectral indices were extracted from the spectral data, and three feature selection methods, recursive feature elimination (RFE), Boruta feature selection, and the Pearson correlation coefficient (PCC), were used to filter high spectral indices in order to reduce the dimensionality of the data. Four major basic learner models, (1) support vector machine (SVM), (2) Gaussian process (GP), (3) linear ridge regression (LRR), and (4) random forest (RF), were also constructed, and an ensemble machine learning model was developed by combining the four base learner models. The results showed that the SVM yield prediction model, constructed on the basis of the preferred features, performed the best among the base learner models, with an R2 between 0.62 and 0.73. The accuracy of the proposed ensemble learner model was higher than that of each base learner model; moreover, the R2 (0.78) for the yield prediction model based on Boruta’s preferred characteristics was the highest at the grain-filling stage.

Aedes aegypti and Aedes albopictus abundance, landscape coverage and spectral indices effects in a subtropical city of Argentina

The presence, abundance and distribution of Aedes (Stegomyia) aegypti (Linnaeus 1762) and Aedes (Stegomyia) albopictus (Skuse 1894) could be conditioned by different data obtained from satellite remote sensors. In this paper, we aim to estimate the effect of landscape coverage and spectral indices on the abundance of Ae. aegypti and Ae. albopictus from the use of satellite remote sensors in Eldorado, Misiones, Argentina. Larvae of Aedes aegypti and Ae. albopictus were collected monthly from June 2016 to April 2018, in four outdoor environments: tire repair shops, cemeteries, family dwellings, and an urban natural park. The proportion of each land cover class was determined by Sentinel-2 image classification. Furthermore spectral indices were calculated. Generalized Linear Mixed Models were developed to analyze the possible effects of landscape coverage and vegetation indices on the abundance of mosquitoes. The model's results showed the abundance of Ae. aegypti was better modeled by the minimum values of the NDVI index, the maximum values of the NDBI index and the interaction between both variables. In contrast, the abundance of Ae. albopictus has to be better explained by the model that includes the variables bare soil, low vegetation and the interaction between both variables.

Variational-Scale Segmentation for Multispectral Remote-Sensing Images Using Spectral Indices

Many studies have focused on performing variational-scale segmentation to represent various geographical objects in high-resolution remote-sensing images. However, it remains a significant challenge to select the most appropriate scales based on the geographical-distribution characteristics of ground objects. In this study, we propose a variational-scale multispectral remote-sensing image segmentation method using spectral indices. Real scenes in remote-sensing images contain different types of land cover with different scales. Therefore, it is difficult to segment images optimally based on the scales of different ground objects. To guarantee image segmentation of ground objects with their own scale information, spectral indices that can be used to enhance some types of land cover, such as green cover and water bodies, were introduced into marker generation for the watershed transformation. First, a vector field model was used to determine the gradient of a multispectral remote-sensing image, and a marker was generated from the gradient. Second, appropriate spectral indices were selected, and the kernel density estimation was used to generate spectral-index marker images based on the analysis of spectral indices. Third, a series of mathematical morphology operations were used to obtain a combined marker image from the gradient and the spectral index markers. Finally, the watershed transformation was used for image segmentation. In a segmentation experiment, an optimal threshold for the spectral-index-marker generation method was identified. Additionally, the influence of the scale parameter was analyzed in a segmentation experiment based on a five-subset dataset. The comparative results for the proposed method, the commonly used watershed segmentation method, and the multiresolution segmentation method demonstrate that the proposed method yielded multispectral remote-sensing images with much better performance than the other methods.

Spectral and Soil Quality Index for Monitoring Environmental Rehabilitation and Soil Carbon Stock in an Amazonian Sandstone Mine

Rehabilitation is the key factor for improving soil quality and soil carbon stock after mining operations. Monitoring is necessary to evaluate the progress of rehabilitation and its success, but the use of repeated field surveys is costly and time-consuming at a large scale. This study aimed to monitor the environmental/soil rehabilitation process of an Amazonian sandstone mine by applying spectral indices for predicting soil organic carbon (SOC) stock and comparing them to soil quality index. The studied area has different chronological rehabilitation stages: initial, intermediate, and advanced with 2, 10, and 12 years of onset rehabilitation activities, respectively. Non-rehabilitated (NR) and two native forest areas (RA) were used as controls. Soil samples were analyzed for physical, chemical, and biological attributes. After determination of Normalized Difference Vegetation Index and Bare Soil Index, simple regression analysis comparing these indices with SOC stock showed a good fit (R2 = 0.82). Rehabilitated areas presented higher soil quality index (~1.50-fold) and SOC stock (~10.6-fold) than NR; however, they did not differ of RA. The use of spectral indices was effective for monitoring the soil quality in this study, with a positive correlation between the predicted SOC stock and the calculated soil quality index.

Application of satellite imagery to update depth-area-volume relationships in reservoirs in the semiarid region of Northeast Brazil

ABSTRACT Reservoirs are the primary source of water supply in the semiarid region of Pernambuco state, Brazil, because of the constant water scarcity affecting this region. Knowledge of the amount of water available is essential for the effective management of water resources. The volume of water stored in the reservoirs is calculated using the depth-area-volume relationship. However, in most reservoirs in the semiarid region, this relationship is currently out of date. Therefore, the objective of this study was to explore the potential and limitations of the application of the ISODATA unsupervised classification method to calculate the depth-area-volume relationships of reservoirs in the semiarid region of Pernambuco, Brazil. The application of the ISODATA method was evaluated in three reservoirs in the state of Pernambuco, i.e., Poço da Cruz, Barra do Juá, and Jucazinho. The results were compared with the updated curves of reservoirs obtained from bathymetry and recent LiDAR surveys. The ISODATA method presented satisfactory results for the three reservoirs analyzed. The mean absolute error of the volume in Poço da Cruz and Barra do Juá was lower than 1% of the maximum capacity. The use of the ISODATA method meant that the surface area underestimation tendency in the Poço da Cruz reservoir was less than when spectral indices were used.

A Model-Based Assessment of Canopy-Scale Primary Productivity for the Baltic Sea Benthic Vegetation Using Environmental Variables and Spectral Indices

This study investigated the potential to predict primary production in benthic ecosystems using meteorological variables and spectral indices. In situ production experiments were carried out during the vegetation season of 2020, wherein the primary production and spectral reflectance of different communities of submerged aquatic vegetation (SAV) were measured and chlorophyll (Chl a+b) concentration was quantified in the laboratory. The reflectance of SAV was measured both in air and underwater. First, in situ reflectance spectra of each SAV class were used to calculate different spectral indices, and then the indices were correlated with Chl a+b. Indices using red and blue band combinations such as 650/450 and 650/480 nm explained the largest part of variability in Chl a+b for datasets measured in air and underwater. Subsequently, the best-performing indices were used in boosted regression trees (BRT) models, together with meteorological data to predict the community photosynthesis of different SAV classes. The predictive power (R2) of production models were very high, estimated at the range of 0.82-0.87. The variable contributing the most to the model description was SAV class, followed in most cases by the water temperature. Nevertheless, the inclusion of spectral indices significantly improved BRT models, often by over 20%, and surprisingly their contribution mostly exceeded that of photosynthetically active radiation.

Detecting the Spatial Variability of Seagrass Meadows and Their Consequences on Associated Macrofauna Benthic Activity Using Novel Drone Technology

Seagrass meadows are undergoing significant decline locally and globally from human and climatic impacts. Seagrass decline also impacts seagrass-dependent macrofauna benthic activity, interrupts their vital linkage with adjacent habitats, and creates broader degradation through the ecosystem. Seagrass variability (gain and loss) is a driver of marine species diversity. Still, our understanding of macrofauna benthic activity distribution and their response to seagrass variability from remotely sensed drone imagery is limited. Hence, it is critical to develop fine-scale seasonal change detection techniques appropriate to the scale of variability that will apply to dynamic marine environments. Therefore, this research tested the performance of the VIS and VIS+NIR sensors from proximal low altitude remotely piloted aircraft system (RPAS) to detect fine-scale seasonal seagrass variability using spectral indices and a supervised machine learning classification technique. Furthermore, this research also attempted to identify and quantify macrofauna benthic activity from their feeding burrows and their response to seagrass variability. The results from VIS (visible spectrum) and VIS+NIR (visible and near-infrared spectrum) sensors produced a 90–98% classification accuracy. This accuracy established that the spectral indices were fundamental in this study to identify and classify seagrass density. The other important finding revealed that seagrass-associated macrofauna benthic activity showed increased or decreased abundance and distribution with seasonal seagrass variability from drone high spatial resolution orthomosaics. These results are important for seagrass conservation because managers can quickly detect fine-scale seasonal changes and take mitigation actions before the decline of this keystone species affects the entire ecosystem. Moreover, proximal low-altitude, remotely sensed time-series seasonal data provided valuable contributions for documenting spatial ecological seasonal change in this dynamic marine environment.