<p>The LiDAR data of the maize used in this research were acquired from different stages, by different sensors, and from different flight heights, which results in different point densities. The ground reference data collected by LiCOR LAI-2200 represented the leaf area index of a two-row plot.</p>

3D point cloud processing has been a critical task due to the increasing demand of a variety of applications such as urban planning and management, as-built mapping of industrial sites, infrastructure monitoring, and road safety inspection. Point clouds are mainly acquired from two sources, laser scanning and optical imaging systems. However, the original point clouds usually do not provide explicit semantic information, and the collected data needs to undergo a sequence of processing steps to derive and extract the required information. Moreover, according to application requirements, the outcomes from the point cloud processing could be different. This dissertation presents two tiers of data processing. The first tier proposes an adaptive data processing framework to deal with multi-source and multi-platform point clouds. The second tier introduces two point clouds processing strategies targeting applications mainly from urban environments and transportation corridors.<div><br></div><div>For the first tier of data processing, the internal characteristics (e.g., noise level and local point density) of data should be considered first since point clouds might come from a variety of sources/platforms. The acquired point clouds may have a large number of points. Data processing (e.g., segmentation) of such large datasets is time-consuming. Hence, to attain high computational efficiency, this dissertation presents a down-sampling approach while considering the internal characteristics of data and maintaining the nature of the local surface. Moreover, point cloud segmentation is one of the essential steps in the initial data processing chain to derive the semantic information and model point clouds. Therefore, a multi-class simultaneous segmentation procedure is proposed to partition point cloud into planar, linear/cylindrical, and rough features. Since segmentation outcomes could suffer from some artifacts, a series of quality control procedures are introduced to evaluate and improve the quality of the results.<br></div><div><br></div><div>For the second tier of data processing, this dissertation focuses on two applications for high human activity areas, urban environments and transportation corridors. For urban environments, a new framework is introduced to generate digital building models with accurate right-angle, multi-orientation, and curved boundary from building hypotheses which are derived from the proposed segmentation approach. For transportation corridors, this dissertation presents an approach to derive accurate lane width estimates using point clouds acquired from a calibrated mobile mapping system. In summary, this dissertation provides two tiers of data processing. The first tier of data processing, adaptive down-sampling and segmentation, can be utilized for all kinds of point clouds. The second tier of data processing aims at digital building model generation and lane width estimation applications.<br></div>

<p>Thermal and multispectral imagery can provide users with insights into the water stress status and evapotranspiration demand of crops. However, traditional platforms, such as satellites, for these thermal and multispectral sensors are limited in their usefulness due to low spatial and temporal resolution. Small unmanned aircraft system (UAS) have the potential to have similar sensors installed and provide canopy temperature and reflectance information at spatial and temporal resolutions more useful for crop management; however, most of the existing research on the calibration or the estimation of water status were established based on the satellite platforms either for the sensors calibration or water status quantification. There is, therefore, a need to develop methods specifically for UAS-mounted sensors. In this research, a pixel-based calibration and an atmospheric correction method based on in-field approximate blackbody sources were developed for an uncooled thermal camera, and the higher accurate vegetative temperature acquired after calibration was used as inputs to an algorithm developed for high-resolution thermal imagery for calculating crop latent heat flux. At last, a thermal index based on the Bowen ratio is proposed to quantify the water deficit stress in a crop field, along with this, a method for plot-level analysis of various vegetation and thermal indices have been demonstrated to illustrate its broad application to genetic selection. The objective was to develop a workflow to use high-resolution thermal and multispectral imagery to derive indices that can quantify crops water status on a plot level which will facilitate the research related to breeding selection.</p> <p>The camera calibration method can effectively reduce the root mean square error (RMSE) and variability of measurements. The pixel-based thermal calibration method presented here was able to reduce the measurement uncertainty across all the pixels in the images, thus improving the accuracy and reducing the between-pixel variability of the measurements. During field calibration, the RMSE values relative to ground reference targets for two flights in 2017 were reduced from 6.36°C to 1.24°C and from 4.56°C to 1.32°C, respectively. The latent heat flux estimation algorithm yields an RMSE of 65.23 W/m² compared with the ground reference data acquired from porometer. The Bowen ratio has a high correlation with drought conditions quantified using the soil moisture index, stomatal conductance, and crop water stress index (CWSI), which indicates the potential of this index to be used as a water deficit stress indicator. The thermal and multispectral indices on a plot level displayed will facilitate the breeding selection.</p>

<p>High-throughput phenotyping has emerged as a powerful approach to expedite crop breeding programs. Modern remote sensing systems, including manned aircraft, unmanned aerial vehicles (UAVs), and terrestrial platforms equipped with multiple sensors, such as RGB cameras, multispectral, hyperspectral, and infrared thermal sensors, as well as light detection and ranging (LiDAR) scanners are now widely used technologies in advancing high throughput phenotyping. These systems can collect high spatial, spectral, and temporal resolution data on various phenotypic traits, such as plant height, canopy cover, and leaf area. Enhancing the capability of utilizing such remote sensing data for automated phenotyping is crucial in advancing crop breeding. This dissertation focuses on developing deep learning and transfer learning methodologies for crop phenotyping using multi-modality remote sensing and environmental data. The techniques address two main areas: multi-temporal/across-field biomass prediction and multi-scale remote sensing data fusion.</p> <p><br></p> <p>Biomass is a plant characteristic that strongly correlates with biofuel production, but is also influenced by genetic and environmental factors. Previous studies have shown that deep learning-based models are effective in predicting end-of-season biomass for a single year and field. This dissertation includes development of transfer learning methodologies for multiyear,</p> <p>across-field biomass prediction. Feature importance analysis was performed to identify and remove redundant features. The proposed model can incorporate high-dimensional genetic marker data, along with other features representing phenotypic information, environmental conditions, or management practices. It can also predict end-of-season biomass using mid-season remote sensing and environmental data to provide early rankings. The framework was evaluated using experimental trials conducted from 2017 to 2021 at the Agronomy Center for Research and Education (ACRE) at Purdue University. The proposed transfer learning techniques effectively selected the most informative training samples in the target domain, resulting in significant improvements in end-of-season yield prediction and ranking. Furthermore, the importance of input remote sensing features was assessed at different growth stages.</p> <p><br></p> <p>Remote sensing technology enables multi-scale, multi-temporal data acquisition. However, to fully exploit the potential of the acquired data, data fusion techniques that leverage the strengths of different sensors and platforms are necessary. In this dissertation, a generative adversarial network (GAN) based multiscale RGB-guided model and domain adaptation framework were developed to enhance the spatial resolution of multispectral images. The model was trained on limited high spatial resolution images from a wheel-based platform and then applied to low spatial resolution images acquired by UAV and airborne platforms.</p> <p>The strategy was tested in two distinct scenarios, sorghum plant breeding, and urban areas, to evaluate its effectiveness.</p>

<p> </p> <p>Modern remote sensing platforms such as unmanned aerial vehicles (UAVs) that can carry a variety of sensors including RGB frame cameras, hyperspectral (HS) line cameras, and LiDAR sensors are commonly used in several application domains. In order to derive accurate products such as point clouds and orthophotos, sensors’ interior and exterior orientation parameters (IOP and EOP) must be established. These parameters are derived/refined in a triangulation framework through minimizing the discrepancy between conjugate features extracted from involved datasets. Existing triangulation approaches are not general enough to deal with varying nature of data from different sensors/platforms acquired in diverse environmental conditions. This research develops a generic triangulation framework that can handle different types of primitives (e.g., point, linear, and/or planar features), and sensing modalities (e.g., RGB cameras, HS cameras, and/or LiDAR sensors) for delivering accurate products under challenging conditions with a primary focus on digital agriculture and stockpile monitoring application domains. </p>

<p>Forest canopies are a critical component of forest ecosystems as they influence many important functions. Specifically, the structure of forest canopies is a driver of the magnitude and rate of these functions. Therefore, being able to accurately measure canopy structure is crucial to ensure ecological models and forest management plans are as robust and efficient as possible. However, canopies are complex and dynamic entities and thus their structure can be challenging to accurately measure. Here we study the feasibility of using lidar to measure forest canopy structure across large spatial extents by investigating the compatibility of aerial and terrestrial lidar systems. Building on known structure-function relationships measured with terrestrial lidar, we establish grounds for scaling these relationships to the aerial scale. This would enable accurate measures of canopy structural complexity to be acquired at landscape and regional scales without the time and labor requirements of terrestrial data collection. Our results illustrate the potential for measures of canopy height, vegetation area, horizontal cover, and canopy roughness to be upscaled. Furthermore, we highlight the benefit of utilizing multivariate measures of canopy structure, and the capacity of lidar to identify forest structural types. Moving forward, lidar is a tool to be utilized in tandem with other technologies to best understand the spatial and temporal dynamics of forests and the influence of physical ecosystem structure. </p>

<div> <div> <p>Global Navigation Satellite System Reflectometry (GNSS-R) has presented a great potential as an important approach for ocean remote sensing. Numerous studies have demonstrated that the shape of a code-correlation waveform of forward-scattered Global Positioning System (GPS) signals may be used to measure ocean surface roughness and related geophysical parameters such as wind speed. Recent experiments have extended the reflectometry technique to transmissions from communication satellites. Due to the high power and frequencies of these signals, they are more sensitive to smaller scale ocean surface features, which makes communication satellites a promising signal of opportunity (SoOp) for ocean remote sensing. Recent advancements in fundamental physics are represented by the new scattering model and bistatic radar function developed by Voronovich and Zavorotny based on the SSA (Small Slope Approximation). This new model allows the partially coherent scattering in low wind conditions to be correctly described, which overcomes the limitations of diffuse scattering inherited in the conventional KA-GO (Kirchhoff Approximation-Geometric Optics) model. Furthermore, exploration and practice using spaceborne platforms have become a primary research focus, which is highlighted by the launch of CYGNSS (Cyclone Global Navigation Satellite System) in 2016. CYGNSS is a NASA (National Aeronautics and Space Administration) Earth Venture Mission consisting of an 8 micro-satellite constellation of GNSS-R instruments designed to observe tropical cyclones.</p><p>However, in spite of the significant achievements made in the past 10 years, there are still a variety of challenges to be addressed currently in the ocean reflectometry field. To begin with, the airborne demonstration experiments conducted previously for S-band reflectometry provided neither sufficient amount of data nor the desired scenarios to assess high wind retrieval performance of S-band signals. The current L-band empirical model function theoretically does not also apply to S-band reflectometry. With respect to scattering models, there have been no results of actual data processing so far to verify the performance of the SSA model, especially on low wind retrievals. Lastly, the conventional model fitting methods for ocean wind retrievals were proposed for airborne missions, and new approaches will need to be developed to satisfy the requirement of spaceborne systems.<br></p><p>The research described in this thesis is mainly focused on the development, application and evaluation of new models and algorithms for ocean wind remote sensing. The first part of the thesis studies the extension of reflectometry methods to the general class of SoOps. The airborne reception of commercial satellite S-band transmissions is demonstrated under both low and high wind speed conditions. As part of this effort, a new S-band geophysical model function (GMF) is developed for ocean wind remote sensing using S-band data collected in the 2014 NOAA (National Oceanic and Atmospheric Administration) hurricane campaign. The second part introduces a dual polarization L- and S-band reflectometry experiment, performed in collaboration with Naval Research Lab (NRL), to retrieve and analyze surface winds and compare the results with CYGNSS satellite retrievals and NOAA data buoy measurements. The problems associated with low wind speed retrieval arising from near specular surface reflections are studied. Results have shown improved wind speed retrieval accuracy using bistatic radar cross section (BRCS) modeled by the SSA when compared with KA-GO, in the cases of low to medium diffuse scattering. The last part focuses on the contributions to the NASA-funded spaceborne CYGNSS project. It shows that the accuracy of CYGNSS ocean wind retrieval is improved by an Extended Kalman Filter (EKF) algorithm. Compared with the baseline observable methods, preliminary results showed promising accuracy improvement when the EKF was applied to actual CYGNSS data.<br><br></p></div></div>

This thesis considers various aspects of the use of ground-based methods and remote sensing of Biological Soil Crusts (BSC). They are mostly distributed in winter rainfall dominated areas such as those at Middleback Field Centre (MFC) in South Australia. They can be used potentially as an indicator of rangeland condition by estimating grazing pressure (trampling). Two BSC based indicators for rangeland condition assessment are species composition and cover. While there is strong agreement that BSC composition is a good indicator, there is less agreement that BSC cover alone is a good indicator. Although BSC have been included in previous remotely-sensed studies, their spectral characteristics, and hence their contributions to remotely-sensed spectral signatures, are not well known. Data collection methods were refined for suitable method selection, stratification and site characterization, and morphological/ functional group classification. Cover data of BSC were collected using a 100 m line-intercept method on the stratified land units and statistical analyses were based on the cover variance analyses. Spectra of BSC groups were collected and characterized for different remote sensing indices. Five grazing gradient models based on collected spectra were developed for the evaluation of BSC effect on remotely-sensed data. Both existing and newly developed remote sensing indices were examined for BSC detection. Sampling for cover of BSC in the field showed that there is indeed a detectable change with distance from water, suggesting that BSC cover can be used as an indicator of rangeland condition, provided that appropriate stratification of the study sites is carried out prior to sampling, and spectral differences in morphological and functional groups are taken into account. Spectral analysis of BSC components showed that different classes of organisms in the crusts have different spectral characteristics, and in particular, that the (commonly-used) perpendicular vegetation index (PD54) is not suitable for detecting BSC. On the other hand, ground-level spectral modelling showed that the Normalized Difference Vegetation Index (NDVI) and Soil Stability Index (SSI) did show a distinguishable contribution from BSC. A procedure for detecting cover of BSC was developed for image taken during the period after an effective rain, in contrast to the normal practice of selecting images of dry surfaces for interpretation. The most suitable intervals appears to be 2-4 days after rain in late autumn, winter and early spring. Of the existing indices, the SSI is the best for estimating cover of BSC from Landsat images. However, eight new indices, specifically designed for detection of BSC were developed during the cource of this work. The best results were obtained for indices using using the middle-infrared bands. These results are promising for application to rangeland monitoring and suggest that BSC cover is an important indicator of rangeland condition if appropriate stratification, classification and data-collection methods are used. The effects of BSC cover on a remotely-sensed method are considerable, and thus they can not be neglected during image interpretation. There are different phenological patterns for BSC, annual and perennial elements, thus there is the possibility for the selection of imagery based on each phenological stage to detect these elements. Application of certain indices such as the PD54 may create mis-estimation of land covers. Although some of the existing and newly developed indices had significant results for BSC cover estimation, there is a requirement for a standalone remotely-sensed method to conclude the best index.

Recently, products of Unmanned Aerial System (UAS) integrated through SIFT algorithm and dense cloud matching using structure from motion has gained prominence with tree-level inventory maintenance in forestry. Various studies have been carried out by using UAS imagery to quantify and map forest structure of simple coniferous stands. However, most of the previous works employ methodologies that require manual inputs and lack of reproducibility to other forest systmes. Manual detection of trees and calculation of their attributes can be a time-consuming and complicated process which can be overcome with an automated technique applied by forest managers and/or landowners is highly desired to take full advantage of the readily available UAS remote sensing images. This study presents a methodology for automated measurements of tree height, crown area and crown diameter of hardwood species using UAS images. Different UAS platforms were employed to gather digital data of two hardwood plantations at Martell, Indiana. The resulting aerial images were used to generate the Digital Surface Model (DSM) and Digital Elevation Model (DEM) for the forest stand from which the Crown Height Model (CHM) was derived. The canopy height model can be inputted to the web platform deployed through shiny server (https://feilab.shinyapps.io/Crown/) to derive individual tree parameters automatically. The results show that this automated method provides a high accuracy in individual tree identification (F-score> 90%) and tree-level measurements (RMSEht<1.2m and RMSEcrn<1m). Moreover, tree-level parameter estimation for 4,600 trees were calculated in less than 30 minutes based on a post-processed DSM from UAS-SfM derived images with minimal manual inputs. This study demonstrates the feasibility of automated inventory and measure of tree-level attributes in hardwood plantations with UAS images.

<p> </p> <p>Leaf Area Index (LAI) is commonly defined as the total area of a leaf per unit area of the ground. LAI is an important variable for characterizing plant canopy related to the interception of solar radiation. Direct measurement of LAI by destructive sampling is tedious, time-consuming, and labor-intensive. With the advance of remote sensing, studies have explored multispectral and hyperspectral remote sensing image data and LiDAR point clouds as individual sources to estimate LAI indirectly. This study investigates the estimation of LAI for maize row crops over the growing season based on features derived from high resolution LiDAR and hyperspectral data acquired simultaneously from a UAV platform. Support Vector Regression (SVR) models are developed using cross validation and evaluated relative to the contribution of the multi-modality remote sensing data. The study is based on data acquired for experiments in plant breeding and evaluation of nitrogen management practice trials conducted at the Agronomy Center for Research and Education (ACRE) in 2021 and 2022, respectively. Reference data for the models were collected using a LI-COR® LAI-2200-C Plant Canopy Analyzer. Including both LiDAR and hyperspectral data sources in the SVR model improved the 𝑅_ref^2 (relative to 1:1 comparison line), RMSE and Relative RMSE (rRMSE) values for both the plant breeding and nitrogen management practice experiments, although incremental gains were small overall. More importantly, it was observed that the contributions of the LiDAR vs hyperspectral inputs to the models also varied throughout the growing season. </p>