A benchmark dataset for canopy crown detection and delineation in co-registered airborne RGB, LiDAR and hyperspectral imagery from the National Ecological Observation Network.
Ontology highlight
ABSTRACT:
SUBMITTER: Weinstein BG
PROVIDER: S-EPMC8282040 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:An international data science challenge, called National Ecological Observatory Network-National Institute of Standards and Technology data science evaluation, was set up in autumn 2017 with the goal to improve the use of remote sensing data in ecological applications. The competition was divided into three tasks: (1) individual tree crown (ITC) delineation, for identifying the location and size of individual trees; (2) alignment between field surveyed trees and ITCs delineated on remote sensing data; and (3) tree species classification. In this paper, the methods and results of team Fondazione Edmund Mach (FEM) are presented. The ITC delineation (Task 1 of the challenge) was done using a region growing method applied to a near-infrared band of the hyperspectral images. The optimization of the parameters of the delineation algorithm was done in a supervised way on the basis of the Jaccard score using the training set provided by the organizers. The alignment (Task 2) between the delineated ITCs and the field surveyed trees was done using the Euclidean distance among the position, the height, and the crown radius of the ITCs and the field surveyed trees. The classification (Task 3) was performed using a support vector machine classifier applied to a selection of the hyperspectral bands and the canopy height model. The selection of the bands was done using the sequential forward floating selection method and the Jeffries Matusita distance. The results of the three tasks were very promising: team FEM ranked first in the data science competition in Task 1 and 2, and second in Task 3. The Jaccard score of the delineated crowns was 0.3402, and the results showed that the proposed approach delineated both small and large crowns. The alignment was correctly done for all the test samples. The classification results were good (overall accuracy of 88.1%, kappa accuracy of 75.7%, and mean class accuracy of 61.5%), although the accuracy was biased toward the most represented species.
| S-EPMC6330952 | biostudies-literature
Project description:Airborne remote sensing offers unprecedented opportunities to efficiently monitor vegetation, but methods to delineate and classify individual plant species using the collected data are still actively being developed and improved. The Integrating Data science with Trees and Remote Sensing (IDTReeS) plant identification competition openly invited scientists to create and compare individual tree mapping methods. Participants were tasked with training taxon identification algorithms based on two sites, to then transfer their methods to a third unseen site, using field-based plant observations in combination with airborne remote sensing image data products from the National Ecological Observatory Network (NEON). These data were captured by a high resolution digital camera sensitive to red, green, blue (RGB) light, hyperspectral imaging spectrometer spanning the visible to shortwave infrared wavelengths, and lidar systems to capture the spectral and structural properties of vegetation. As participants in the IDTReeS competition, we developed a two-stage deep learning approach to integrate NEON remote sensing data from all three sensors and classify individual plant species and genera. The first stage was a convolutional neural network that generates taxon probabilities from RGB images, and the second stage was a fusion neural network that "learns" how to combine these probabilities with hyperspectral and lidar data. Our two-stage approach leverages the ability of neural networks to flexibly and automatically extract descriptive features from complex image data with high dimensionality. Our method achieved an overall classification accuracy of 0.51 based on the training set, and 0.32 based on the test set which contained data from an unseen site with unknown taxa classes. Although transferability of classification algorithms to unseen sites with unknown species and genus classes proved to be a challenging task, developing methods with openly available NEON data that will be collected in a standardized format for 30 years allows for continual improvements and major gains for members of the computational ecology community. We outline promising directions related to data preparation and processing techniques for further investigation, and provide our code to contribute to open reproducible science efforts.
| S-EPMC8325917 | biostudies-literature
Project description:Global crop production is facing the challenge of a high projected demand, while the yields of major crops are not increasing at sufficient speeds. Crop breeding is an important way to boost crop productivity, however its improvement rate is partially hindered by the long crop generation cycles. If end-season crop traits such as yield can be predicted through early-season phenotypic measurements, crop selection can potentially be made before a full crop generation cycle finishes. This study explored the possibility of predicting soybean end-season traits through the color and texture features of early-season canopy images. Six thousand three hundred and eighty-three images were captured at V4/V5 growth stage over 6039 soybean plots growing at four locations. One hundred and forty color features and 315 gray-level co-occurrence matrix-based texture features were derived from each image. Another two variables were also introduced to account for location and timing differences between the images. Five regression and five classification techniques were explored. Best results were obtained using all 457 predictor variables, with Cubist as the regression technique and Random Forests as the classification technique. Yield (RMSE?=?9.82, R2?=?0.68), Maturity (RMSE?=?3.70, R2?=?0.76) and Seed Size (RMSE?=?1.63, R2?=?0.53) were identified as potential soybean traits that might be early predictable.
| S-EPMC6773688 | biostudies-literature
Project description:In recent years, LIght Detection And Ranging (LiDAR) and especially Terrestrial Laser Scanning (TLS) systems have shown the potential to revolutionise forest structural characterisation by providing unprecedented 3D data. However, manned Airborne Laser Scanning (ALS) requires costly campaigns and produces relatively low point density, while TLS is labour intense and time demanding. Unmanned Aerial Vehicle (UAV)-borne laser scanning can be the way in between. In this study, we present first results and experiences with the RIEGL RiCOPTER with VUX ® -1UAV ALS system and compare it with the well tested RIEGL VZ-400 TLS system. We scanned the same forest plots with both systems over the course of two days. We derived Digital Terrain Model (DTMs), Digital Surface Model (DSMs) and finally Canopy Height Model (CHMs) from the resulting point clouds. ALS CHMs were on average 11.5 c m higher in five plots with different canopy conditions. This showed that TLS could not always detect the top of canopy. Moreover, we extracted trunk segments of 58 trees for ALS and TLS simultaneously, of which 39 could be used to model Diameter at Breast Height (DBH). ALS DBH showed a high agreement with TLS DBH with a correlation coefficient of 0.98 and root mean square error of 4.24 c m . We conclude that RiCOPTER has the potential to perform comparable to TLS for estimating forest canopy height and DBH under the studied forest conditions. Further research should be directed to testing UAV-borne LiDAR for explicit 3D modelling of whole trees to estimate tree volume and subsequently Above-Ground Biomass (AGB).
| S-EPMC5677400 | biostudies-literature
Project description:We report large-scale estimates of Amazonian gap dynamics using a novel approach with large datasets of airborne light detection and ranging (lidar), including five multi-temporal and 610 single-date lidar datasets. Specifically, we (1) compared the fixed height and relative height methods for gap delineation and established a relationship between static and dynamic gaps (newly created gaps); (2) explored potential environmental/climate drivers explaining gap occurrence using generalized linear models; and (3) cross-related our findings to mortality estimates from 181 field plots. Our findings suggest that static gaps are significantly correlated to dynamic gaps and can inform about structural changes in the forest canopy. Moreover, the relative height outperformed the fixed height method for gap delineation. Well-defined and consistent spatial patterns of dynamic gaps were found over the Amazon, while also revealing the dynamics of areas never sampled in the field. The predominant pattern indicates 20-35% higher gap dynamics at the west and southeast than at the central-east and north. These estimates were notably consistent with field mortality patterns, but they showed 60% lower magnitude likely due to the predominant detection of the broken/uprooted mode of death. While topographic predictors did not explain gap occurrence, the water deficit, soil fertility, forest flooding and degradation were key drivers of gap variability at the regional scale. These findings highlight the importance of lidar in providing opportunities for large-scale gap dynamics and tree mortality monitoring over the Amazon.
| S-EPMC7809196 | biostudies-literature
Project description:Understanding the carbon flux of forests is critical for constraining the global carbon cycle and managing forests to mitigate climate change. Monitoring forest growth and mortality rates is critical to this effort, but has been limited in the past, with estimates relying primarily on field surveys. Advances in remote sensing enable the potential to monitor tree growth and mortality across landscapes. This work presents an approach to measure tree growth and loss using multidate lidar campaigns in a high-biomass forest in California, USA. Individual tree crowns were delineated in 2008 and again in 2013 using a 3D crown segmentation algorithm, with derived heights and crown radii extracted and used to estimate individual tree aboveground biomass. Tree growth, loss, and aboveground biomass were analyzed with respect to tree height and crown radius. Both tree growth and loss rates decrease with increasing tree height, following the expectation that trees slow in growth rate as they age. Additionally, our aboveground biomass analysis suggests that, while the system is a net source of aboveground carbon, these carbon dynamics are governed by size class with the largest sources coming from the loss of a relatively small number of large individuals. This study demonstrates that monitoring individual tree-based growth and loss can be conducted with multidate airborne lidar, but these methods remain relatively immature. Disparities between lidar acquisitions were particularly difficult to overcome and decreased the sample of trees analyzed for growth rate in this study to 21% of the full number of delineated crowns. However, this study illuminates the potential of airborne remote sensing for ecologically meaningful forest monitoring at an individual tree level. As methods continue to improve, airborne multidate lidar will enable a richer understanding of the drivers of tree growth, loss, and aboveground carbon flux.
| S-EPMC5980410 | biostudies-literature
Project description:With the advent of Sentinel-2, it is now possible to generate large-scale chlorophyll content maps with unprecedented spatial and temporal resolution, suitable for monitoring ecological processes such as vegetative stress and/or decline. However methodological gaps exist for adapting this technology to heterogeneous natural vegetation and for transferring it among vegetation species or plan functional types. In this study, we investigated the use of Sentinel-2A imagery for estimating needle chlorophyll (Ca+b) in a sparse pine forest undergoing significant needle loss and tree mortality. Sentinel-2A scenes were acquired under two extreme viewing geometries (June vs. December 2016) coincident with the acquisition of high-spatial resolution hyperspectral imagery, and field measurements of needle chlorophyll content and crown leaf area index. Using the high-resolution hyperspectral scenes acquired over 61 validation sites we found the CI chlorophyll index R750/R710 and Macc index (which uses spectral bands centered at 680?nm, 710?nm and 780?nm) had the strongest relationship with needle chlorophyll content from individual tree crowns (r2?=?0.61 and r2?=?0.59, respectively; p?<?0.001), while TCARI and TCARI/OSAVI, originally designed for uniform agricultural canopies, did not perform as well (r2?=?0.21 and r2?=?0.01, respectively). Using lower-resolution Sentinel-2A data validated against hyperspectral estimates and ground truth needle chlorophyll content, the red-edge index CI and the Sentinel-specific chlorophyll indices CI-Gitelson, NDRE1 and NDRE2 had the highest accuracy (with r2 values >0.7 for June and >0.4 for December; p?<?0.001). The retrieval of needle chlorophyll content from the entire Sentinel-2A bandset using the radiative transfer model INFORM yielded r2?=?0.71 (RMSE?=?8.1??g/cm2) for June, r2?=?0.42 (RMSE?=?12.2??g/cm2) for December, and r2?=?0.6 (RMSE?=?10.5??g/cm2) as overall performance using the June and December datasets together. This study demonstrates the retrieval of leaf Ca+b with Sentinel-2A imagery by red-edge indices and by an inversion method based on a hybrid canopy reflectance model that accounts for tree density, background and shadow components common in sparse forest canopies.
| S-EPMC6472624 | biostudies-literature
Project description:Above-ground biomass (AGB) is a trait with much potential for exploitation within wheat breeding programs and is linked closely to canopy height (CH). However, collecting phenotypic data for AGB and CH within breeding programs is labor intensive, and in the case of AGB, destructive and prone to assessment error. As a result, measuring these traits is seldom a priority for breeders, especially at the early stages of a selection program. LiDAR has been demonstrated as a sensor capable of collecting three-dimensional data from wheat field trials, and potentially suitable for providing objective, non-destructive, high-throughput estimates of AGB and CH for use by wheat breeders. The current study investigates the deployment of a LiDAR system on a ground-based high-throughput phenotyping platform in eight wheat field trials across southern Australia, for the non-destructive estimate of AGB and CH. LiDAR-derived measurements were compared to manual measurements of AGB and CH collected at each site and assessed for their suitability of application within a breeding program. Correlations between AGB and LiDAR Projected Volume (LPV) were generally strong (up to r = 0.86), as were correlations between CH and LiDAR Canopy Height (LCH) (up to r = 0.94). Heritability (H2) of LPV (H2 = 0.32-0.90) was observed to be greater than, or similar to, the heritability of AGB (H2 = 0.12-0.78) for the majority of measurements. A similar level of heritability was observed for LCH (H2 = 0.41-0.98) and CH (H2 = 0.49-0.98). Further to this, measurements of LPV and LCH were shown to be highly repeatable when collected from either the same or opposite direction of travel. LiDAR scans were collected at a rate of 2,400 plots per hour, with the potential to further increase throughput to 7,400 plots per hour. This research demonstrates the capability of LiDAR sensors to collect high-quality, non-destructive, repeatable measurements of AGB and CH suitable for use within both breeding and research programs.
| S-EPMC6775483 | biostudies-literature
Project description:Producing quantitative and reliable measures of crop disease is essential for resistance breeding, but is challenging and time consuming using traditional phenotyping methods. Hyperspectral remote sensing has shown potential for the detection of plant diseases, but its utility for phenotyping large and diverse populations of plants under field conditions requires further evaluation. In this study, we collected canopy hyperspectral data from 335 wheat varieties using a spectroradiometer, and we investigated the use of canopy reflectance for detecting the Septoria tritici blotch (STB) disease and for quantifying the severity of infection. Canopy- and leaf-level infection metrics of STB based on traditional visual assessments and automated analyses of leaf images were used as ground truth data. Results showed (i) that canopy reflectance and the selected spectral indices show promise for quantifying STB infections, and (ii) that the normalized difference water index (NDWI) showed the best performance in detecting STB compared to other spectral indices. Moreover, partial least squares (PLS) regression models allowed for an improvement in the prediction of STB metrics. The PLS discriminant analysis (PLSDA) model calibrated based on the spectral data of four reference varieties was able to discriminate between the diseased and healthy canopies among the 335 varieties with an accuracy of 93% (Kappa = 0.60). Finally, the PLSDA model predictions allowed for the identification of wheat genotypes that are potentially more susceptible to STB, which was confirmed by the STB visual assessment. This study demonstrates the great potential of using canopy hyperspectral remote sensing to improve foliar disease assessment and to facilitate plant breeding for disease resistance.
| S-EPMC6108383 | biostudies-literature
Project description:RGB displays effectively simulate millions of colors in the eyes of humans by modulating the relative amount of light emitted by 3 differently colored juxtaposed lights (red, green, and blue). The relationship between the ratio of red, green, and blue light and the perceptual experience of that light has been well defined by psychophysical experiments in humans, but is unknown in animals. The perceptual experience of an animal looking at an RGB display of imagery designed for humans is likely to poorly represent an animal's experience of the same stimulus in the real world. This is due, in part, to the fact that many animals have different numbers of photoreceptor classes than humans do and that their photoreceptor classes have peak sensitivities centered over different parts of the ultraviolet and visible spectrum. However, it is sometimes possible to generate videos that accurately mimic natural stimuli in the eyes of another animal, even if that animal's sensitivity extends into the ultraviolet portion of the spectrum. How independently each RGB phosphor stimulates each of an animal's photoreceptor classes determines the range of colors that can be simulated for that animal. What is required to determine optimal color rendering for another animal is a device capable of measuring absolute or relative quanta of light across the portion of the spectrum visible to the animal (i.e., a spectrometer), and data on the spectral sensitivities of the animal's photoreceptor classes. In this article, we outline how to use such equipment and information to generate video stimuli that mimic, as closely as possible, an animal's color perceptual experience of real-world objects. Key words: color vision, computer animation, perception, video playback, virtual reality.
| S-EPMC5804149 | biostudies-literature