Papers by Gilbert Kattenborn
ISPRS Annals of Photogrammetry, Remote Sensing and Spatial Information Sciences, 2018
One fundamental metric to characterize trees and forest stands is the diameter at breast height (... more One fundamental metric to characterize trees and forest stands is the diameter at breast height (DBH). However, the vertical geometry of tree stems hampers a direct measurement by means of orthographic aerial imagery. Nevertheless, the DBH in deciduous forest stands could be measured from UAV-based imagery using the width of a stem´s cast shadow projected on the ground. Here, we compare in-situ measured DBH of 100 trees with the DBH visually interpreted from cast-shadows derived in UAV-based aerial imagery. Then, based on simulated datasets, we determine suitable DBH sampling sizes for a robust and efficient retrieval of stand diameter distributions. The UAV-based DBH estimation resulted in an r² of 0.74, RMSE of 7.61 cm, NRMSE of 12.8% and approximately unbiased results. According to our simulations it can be assumed that a sample size of 25-50 individual DBH measurements per forest stand allows estimating reliable diameter distributions. The presented pilot study gives a first insight on the potential of such an approach for operational assessments of diameter distribution in deciduous forest stands and might be particularly interesting for stands in difficult terrain situations. The presented approach can be extended to estimate the basal area, timber stock or biomass.
IEEE 1999 International Geoscience and Remote Sensing Symposium. IGARSS'99 (Cat. No.99CH36293), 1999
Singularity and non-stationarity in images, such as edges and contours in the deterministic case,... more Singularity and non-stationarity in images, such as edges and contours in the deterministic case, or fractal processes in the stochastic case, carry important thematic information. In particular we are interested in the evolution of structures with scale and polarization state in SAR images for land applications. We propose a technique to extract this type of information based on a multi-resolution
Geocarto International, 1997
ABSTRACT A method to correct atmospheric effects on Landsat‐TM data over mountainous terrain in t... more ABSTRACT A method to correct atmospheric effects on Landsat‐TM data over mountainous terrain in the solar spectrum was developed primarily for forestry applications, but can also be adapted to other sensors and applications. The algorithm automatically identifies patches of dark, dense vegetation as controllable dark surfaces for the purpose of estimating aerosol optical thickness. Thereby a network of radiometric control points, rendering both horizontal as well as relief induced vertical changes of atmospheric conditions, is expanded over a scene. Atmospheric parameters calculated with LOWTRAN over these pixels are smoothed and interpolated by cubic spline regression. With these values atmospherically corrected ground reflectances are calculated pixelwise by iteration of a correction model. Improvements in visual interpretation, in the standardization of photo products and in the use of calibrated data with multitemporal applications are discussed.
Geocarto International, 1997
ABSTRACT A method to correct atmospheric effects on Landsat‐TM data over mountainous terrain in t... more ABSTRACT A method to correct atmospheric effects on Landsat‐TM data over mountainous terrain in the solar spectrum was developed primarily for forestry applications, but can also be adapted to other sensors and applications. The algorithm automatically identifies patches of dark, dense vegetation as controllable dark surfaces for the purpose of estimating aerosol optical thickness. Thereby a network of radiometric control points, rendering both horizontal as well as relief induced vertical changes of atmospheric conditions, is expanded over a scene. Atmospheric parameters calculated with LOWTRAN over these pixels are smoothed and interpolated by cubic spline regression. With these values atmospherically corrected ground reflectances are calculated pixelwise by iteration of a correction model. Improvements in visual interpretation, in the standardization of photo products and in the use of calibrated data with multitemporal applications are discussed.
One fundamental metric to characterize trees and forest stands is the diameter at breast height (... more One fundamental metric to characterize trees and forest stands is the diameter at breast height (DBH). However, the vertical geometry of tree stems hampers a direct measurement by means of orthographic aerial imagery. Nevertheless, the DBH in deciduous forest stands could be measured from UAV-based imagery using the width of a stem´s cast shadow projected on the ground. Here, we compare in-situ measured DBH of 100 trees with the DBH visually interpreted from cast-shadows derived in UAV-based aerial imagery. Then, based on simulated datasets, we determine suitable DBH sampling sizes for a robust and efficient retrieval of stand diameter distributions. The UAV-based DBH estimation resulted in an r² of 0.74, RMSE of 7.61 cm, NRMSE of 12.8% and approximately unbiased results. According to our simulations it can be assumed that a sample size of 25-50 individual DBH measurements per forest stand allows estimating reliable diameter distributions. The presented pilot study gives a first insight on the potential of such an approach for operational assessments of diameter distribution in deciduous forest stands and might be particularly interesting for stands in difficult terrain situations. The presented approach can be extended to estimate the basal area, timber stock or biomass.
Proceedings of IEEE Topical Symposium on Combined Optical, Microwave, Earth and Atmosphere Sensing, 1993
ABSTRACT Not Available
Proceedings of IGARSS '93 - IEEE International Geoscience and Remote Sensing Symposium, 1993
ABSTRACT The spatial statistics in a SAR (synthetic aperture radar) image is an important paramet... more ABSTRACT The spatial statistics in a SAR (synthetic aperture radar) image is an important parameter for the classification of different targets, in two respects: it can be an additional discriminant and it can be used as the basis in speckle filtering techniques. The authors survey some statistical image models presented in the literature, and then focus on the experimental characterization of the spatial statistics in SAR imagery. Experimental data acquired by the JPL AIRSAR multifrequency polarimetric SAR sensor within the MAESTRO-1 and MAC-Europe campaign, as well as the ESA ERS-1 data are used in the study. MAESTRO 1 was a campaign flown in 1989 and jointly funded by ESA and the CEC JRC Ispra. MAC was a multisensor airborne campaign organized by NASA in 1991. The authors characterize texture in an intensity image by 2 parameters (first order and second order statistics): the alpha parameter in the K distribution, the autocorrelation function. Methods are presented to estimate experimentally these parameters. The parameters are studied for different types of targets and as a function of other parameters, such as polarization, frequency band, viewing direction, multilook processing, spatial resolution
IGARSS'97. 1997 IEEE International Geoscience and Remote Sensing Symposium Proceedings. Remote Sensing - A Scientific Vision for Sustainable Development, 1997
ERS-1 acquisitions of the Black Forest test site in Germany from 1991 to 1993 were selected to co... more ERS-1 acquisitions of the Black Forest test site in Germany from 1991 to 1993 were selected to cover a variety of seasonal and phenological stages. Multisensor analysis (ERS-1, SPOT) was performed using a forestry GIS data base from the regular forest taxation. It turned out that the C-VV band of ERS-1 is sensitive to age in young deciduous and coniferous forest stands of temperate zones. A synergetic relationship between optical and microwave data has been established. The combined data set provided a combination that allowed better differentiation of forest age classes, as compared to results obtained with the individual sensors
The measurement of geometric tree attributes is a major part of forest inventories. In practice r... more The measurement of geometric tree attributes is a major part of forest inventories. In practice respective attributes are mostly measured by terrestrial surveys, less often by laser scanning. These methods are relatively time consuming or costly. The presented study was conducted in order to evaluate photogrammetric point clouds, based on Unmanned Aerial Vehicles (UAV) photo flights, as a cost effective and fast additional instrument for single tree detection and tree height measurements within forest stands. A coniferous and a deciduous forest site near Freiburg im Breisgau, Germany were surveyed during the leaf-on state. For each site point clouds were generated photogrammetrically from UAV-based aerial photographs. The point clouds were processed with software developed for analyzing Light Detection And Ranging (LiDAR) data for forestry applications, to generate virtual tree models. Multiple models were calculated, by varying processing parameters, and analyzed for their accuracy...
Proceedings of IGARSS '94 - 1994 IEEE International Geoscience and Remote Sensing Symposium, 1994
ABSTRACT It is among the basic principles of physical measurement that changing the conditions of... more ABSTRACT It is among the basic principles of physical measurement that changing the conditions of measure results in improving the knowledge of the observed phenomena of interest. In remote sensing this is achieved in the case of multispectral optical data using wavelength diversity. In the the case of single band ERS-1 data the action of external factors (meteorological, phenological, land practices) have an analogous discriminant effect considering multitemporal data. Improvement of extraction of agricultural information is performed using appropriate preprocessing, standardized principal component analysis, and classification using maximum entropy based method, in order to estimate crops acreage. The experiment, conducted over the great Driffield MARS project test site (ERS-1 data from November 1992 to may 1993) shows that set aside land can be identified as early as in February, and that acreage estimation can be done in may for the most important crop types
Proceedings of IGARSS '94 - 1994 IEEE International Geoscience and Remote Sensing Symposium, 1994
ABSTRACT A theoretical microwave backscatter model was used to simulate the sensitivity of ERS-1 ... more ABSTRACT A theoretical microwave backscatter model was used to simulate the sensitivity of ERS-1 SAR data over temperate forest to SAR system, forest stand, and environmental parameters. Of particular interest were the simulations of the temperature dependence of the backscattering coefficient under consideration of the partial freezing of the soil and plant water at temperatures below 0°C and of leaf and needle biomass variations as they occur due to the seasonal development and as a consequence of stress and disease
IEEE 1999 International Geoscience and Remote Sensing Symposium. IGARSS'99 (Cat. No.99CH36293), 1999
Singularity and non-stationarity in images, such as edges and contours in the deterministic case,... more Singularity and non-stationarity in images, such as edges and contours in the deterministic case, or fractal processes in the stochastic case, carry important thematic information. In particular we are interested in the evolution of structures with scale and polarization state in SAR images for land applications. We propose a technique to extract this type of information based on a multi-resolution
Geocarto International, 1997
ABSTRACT A method to correct atmospheric effects on Landsat‐TM data over mountainous terrain in t... more ABSTRACT A method to correct atmospheric effects on Landsat‐TM data over mountainous terrain in the solar spectrum was developed primarily for forestry applications, but can also be adapted to other sensors and applications. The algorithm automatically identifies patches of dark, dense vegetation as controllable dark surfaces for the purpose of estimating aerosol optical thickness. Thereby a network of radiometric control points, rendering both horizontal as well as relief induced vertical changes of atmospheric conditions, is expanded over a scene. Atmospheric parameters calculated with LOWTRAN over these pixels are smoothed and interpolated by cubic spline regression. With these values atmospherically corrected ground reflectances are calculated pixelwise by iteration of a correction model. Improvements in visual interpretation, in the standardization of photo products and in the use of calibrated data with multitemporal applications are discussed.
Uploads
Papers by Gilbert Kattenborn