Evaluation of simulated bands in airborne optical sensors for tree species identification

Pant, Paras; Heikkinen, Ville; Hovi, Aarne; Korpela, Ilkka; Hauta-Kasari, Markku; Tokola, Timo

doi:10.1016/j.rse.2013.07.016

Cited by 26 publications

(29 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The potential of ADS40 sensors in tree species classification can be improved by an additional band in the red-edge, λ = 700 nm (Heikkinen et al 2010;Pant et al 2013). The most recent version of the ADS40 is a three-line four-band sensor.…”

Section: Practical Implications and Suggestions For Future Researchmentioning

confidence: 99%

“…Tests using smaller feature sets were not reported, however this constitutes an interesting research question. Pant et al (2013) showed that an optimized five-band (wide-band) sensor provides stand-level data that result in comparable classification performance with a HS sensor. The geometry is a particular issue in HS imaging, because the exterior orientation is exclusively dependent on direct georeferencing, and the interior orientation (passage of image rays from the projection center to the pixels of different bands) is not as accurate as in photogrammetric sensors.…”

Section: Practical Implications and Suggestions For Future Researchmentioning

confidence: 99%

“…Thus, there can be mismatch between the HS and other data, which may hinder automatic analyses and data fusion (e.g. Dalponte et al 2013Dalponte et al , 2014Pant et al 2013).…”

Section: Practical Implications and Suggestions For Future Researchmentioning

confidence: 99%

See 2 more Smart Citations

Tree species identification in aerial image data using directional reflectance signatures

Korpela¹,

Mehtätalo²,

Markelin³

et al. 2014

Silva Fenn.

Self Cite

View full text Add to dashboard Cite

Tree species identification in aerial image data using directional reflectance signatures Korpela I., Mehtätalo L., Markelin L., Seppänen A., Kangas A. (2014). Tree species identification in aerial image data using directional reflectance signatures. Silva Fennica vol. 48 no. 3 article id 1087. 20 p. Highlights• Multispectral reflectance data showed a strong and spectrally correlated tree effect.• There was no gain in species classification from using species-specific differences of directional reflectance in real data and only a marginal improvement in simulated data.• The directional signatures extracted in multiple images are obscured by the intrinsic withinspecies variation, correlated observations and inherent reflectance calibration errors. AbstractTree species identification using optical remote sensing is challenging. Modern digital photogrammetric cameras enable radiometrically quantitative remote sensing and the estimation of reflectance images, in which the observations depend largely on the reflectance properties of targets. Previous research has shown that there are species-specific differences in how the brightness observed changes when the viewing direction in an aerial image is altered. We investigated if accounting for such directional signatures enhances species classification, using atmospherically corrected, real and simulated multispectral Leica ADS40 line-camera data. Canopy in direct and diffuse illumination were differentiated and species-specific variance-covariance structures were analyzed in real reflectance data, using mixed-effects modeling. Species classification simulations aimed at elucidating the level of accuracy that can be achieved by using images of different quality, number and view-illumination geometry. In real data, a substantial variance component was explained by tree effect, which demonstrates that observations from a tree correlate between observation geometries as well as spectrally. Near-infrared band showed the strongest tree effect, while the directionality was weak in that band. The gain from directional signatures was insignificant in real data, while simulations showed a potential gain of 1-3 percentage points in species classification accuracy. The quality of reflectance calibration was found to be important as well as the image acquisition geometry. We conclude that increasing the number of image observations cancels out random observation noise and reflectance calibration errors, but fails to eliminate the tree effect and systematic calibration inaccuracy. Directional reflectance constitutes a marginal improvement in tree species classification.

show abstract

Section: Practical Implications and Suggestions For Future Researchmentioning

confidence: 99%

Section: Practical Implications and Suggestions For Future Researchmentioning

confidence: 99%

See 1 more Smart Citation

Tree species identification in aerial image data using directional reflectance signatures

Korpela¹,

Mehtätalo²,

Markelin³

et al. 2014

Silva Fenn.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Tree heights can be measured using specific remote sensing sensors like LiDAR, interferometric synthetic aperture radar (InSAR), or by photogrammetry, while tree diameter requires field measurements, and species can partly be estimated in boreal areas using near infrared observations of crown shape, e.g. [2]. Although homogenous forest stand can be mapped with optical data, the species recognition is very difficult task in mixed-species stands and tropical areas.…”

Section: Introductionmentioning

confidence: 99%

Remote Sensing Concepts and Their Applicability in REDD+ Monitoring

Tokola

2015

Curr Forestry Rep

Self Cite

View full text Add to dashboard Cite

Th e Tier app ro ach introd uced by the ntergovernmental Panel on Climate Change (IPCC) has been used to describe levels of methodological quality and complexity in the measurement reporting and verification (MRV) assessment system. Although guidelines do not explicitly state accuracy requirements, it has been asserted in previous studies that remote sensing missions should entail biomass errors that remain within 20 % of stand level estimates. The biomass of the target forest can be almost directly measured using LiDAR techniques when other sensors have rather low capacity to penetrate through canopy and response signals are not too noisy to detect changes in closed forests. Optical or radar satellite data, in relative terms, are only a tenth of the price of LiDAR. Taking into account cost-effectiveness within a tropical context, it is proposed that Landsat-like data can be useful for obtaining land cover and economical estimates for large areas. Typically, the role of airborne LiDAR is to provide ground truth type of information covering large areas. The reducing emissions from deforestation and forest degradation measurement, reporting and verification (REDD MRV) will need good ground sample for basal area and canopy cover change detection. LiDAR will be also feasible for forest management planning maps and planning operational REDD activities, but cost-effective mapping of carbon stocks for regional-and national-scale assessments will rely on a combination of satellite imagery and ground-based inventory samples. The distribution of tree species and different forest types need to be measured from the field and information can be generalised using satellite data with 1-30-m resolution. The global scale maps with the ground resolution 100 m-1 km are mainly for visualisation purposes and can be used to extrapolate distribution of information collected by 1-30-m remote sensing sensors and field plots.

show abstract

“…Tree species classification is one important application for these data. Tree species classification in boreal forests using airborne hyperspectral measurements has been previously considered [2], [3]. A disadvantage of the HSI is that the spatial resolution is coarser than in operational photogrammetric multispectral images.…”

Section: Introductionmentioning

confidence: 99%

Logistic Regression-Based Spectral Band Selection for Tree Species Classification: Effects of Spatial Scale and Balance in Training Samples

Pant

Heikkinen

Korpela

et al. 2014

IEEE Geosci. Remote Sensing Lett.

Self Cite

View full text Add to dashboard Cite

In this letter, we evaluated the pixel-level and plotlevel tree species classification of Scots Pine, Norway Spruce, and deciduous birch in a boreal forest using 64-band AisaEAGLE II hyperspectral data in a wavelength range of 400-1000 nm. First, band selection was performed using a sparse logistic regressionbased feature selection algorithm with pixel-level and plot-level data in case of balanced and imbalanced training data. This resulted in 8-11 selected hyperspectral bands, depending on the properties of the data used. We evaluated a tree species classification with 8-11 selected hyperspectral bands directly for a least squares support vector machine (LS-SVM)-based pixel-level classification with a relatively small training set size (0.5%-1.5% of the total data) and obtained an accuracy and kappa of around 93.50% and 0.90, respectively. These results are around 0.53%-0.94% points lower than those obtained using all of the hyperspectral bands. Second, one important wavelength region highlight by the selected bands was used to modify the sensor sensitivity configuration in the Leica Airborne Digital Sensor 40 (ADS40) multispectral sensor. Using a simulation model and the hyperspectral data, the modified and standard Leica ADS40 sensor responses were simulated and compared, and the modified system simulated response indicates a 3%-5% point improvement in the pixel-level and plot-level LS-SVM classification accuracy compared with the simulated responses of the standard Leica ADS40 band configuration.

show abstract

Evaluation of simulated bands in airborne optical sensors for tree species identification

Cited by 26 publications

References 26 publications

Tree species identification in aerial image data using directional reflectance signatures

Tree species identification in aerial image data using directional reflectance signatures

Remote Sensing Concepts and Their Applicability in REDD+ Monitoring

Logistic Regression-Based Spectral Band Selection for Tree Species Classification: Effects of Spatial Scale and Balance in Training Samples

Contact Info

Product

Resources

About