Leaf normal distribution is an important structural characteristic of the forest canopy. Although terrestrial laser scanners（TLS） have potential for estimating canopy structural parameters, distinguishing between le...Leaf normal distribution is an important structural characteristic of the forest canopy. Although terrestrial laser scanners（TLS） have potential for estimating canopy structural parameters, distinguishing between leaves and nonphotosynthetic structures to retrieve the leaf normal has been challenging. We used here an approach to accurately retrieve the leaf normals of camphorwood（Cinnamomum camphora） using TLS point cloud data.First, nonphotosynthetic structures were filtered by using the curvature threshold of each point. Then, the point cloud data were segmented by a voxel method and clustered by a Gaussian mixture model in each voxel. Finally, the normal vector of each cluster was computed by principal component analysis to obtain the leaf normal distribution. We collected leaf inclination angles and estimated the distribution, which we compared with the retrieved leaf normal distribution. The correlation coefficient between measurements and obtained results was 0.96, indicating a good coincidence.展开更多
One of the major limitations of using Interferometric Synthetic Aperture Radar(InSAR)in time series analysis is the low-phase coherence associated with rough terrain and vegetated areas,which results in limited spatia...One of the major limitations of using Interferometric Synthetic Aperture Radar(InSAR)in time series analysis is the low-phase coherence associated with rough terrain and vegetated areas,which results in limited spatial coverage in such regions.Permanent scatterers technique was introduced to overcome this limitation using time-series analysis.However,identifying major scatterers within a pixel requires the single-looked pixels oversampling which can be a demanding process especially with large interferometric stacks and vast study areas.Therefore,using multilooked temporal coherent pixels was proposed to increase processing efficiency and coverage by utilizing distributed targets,but this technique may exclude pixels with reliable phase returns because of their temporal varying neighboring pixels.In this paper,we propose a technique to identify multilooked temporal stable pixels with reliable phase returns independent of their neighboring pixels.We conduct a simulation analysis to relate the spatial coherence of a pixel with its expected temporal correlation in the time series analysis module.We found that a liberal temporal correlation threshold of 0.53 in multilooked pixels stack is equivalent to a spatial coherence threshold of 0.2 when using number of looks of 9,which is considered acceptable in temporal coherent pixels,in terms of phase standard deviation.Applying these findings to study the 2011 Tohoku earthquake in the northeastern part of Japan resulted in increasing the number of usable pixels and spatial coverage index by nearly 50.4%and 36.8%,respectively,compared to the temporal coherent pixels.Furthermore,we propose an approach to integrate GPS observations with InSAR time series analysis,which resulted in deformation maps of the megathrust 2011 Tohoku earthquake with mean RMSE of 11.4 mm and a correlation of 98%in comparison to GPS observations.展开更多
文摘Leaf normal distribution is an important structural characteristic of the forest canopy. Although terrestrial laser scanners（TLS） have potential for estimating canopy structural parameters, distinguishing between leaves and nonphotosynthetic structures to retrieve the leaf normal has been challenging. We used here an approach to accurately retrieve the leaf normals of camphorwood（Cinnamomum camphora） using TLS point cloud data.First, nonphotosynthetic structures were filtered by using the curvature threshold of each point. Then, the point cloud data were segmented by a voxel method and clustered by a Gaussian mixture model in each voxel. Finally, the normal vector of each cluster was computed by principal component analysis to obtain the leaf normal distribution. We collected leaf inclination angles and estimated the distribution, which we compared with the retrieved leaf normal distribution. The correlation coefficient between measurements and obtained results was 0.96, indicating a good coincidence.
基金supported by Japanese Government(Monbukagakusho,MEXT)Scholarship in 2012.
文摘One of the major limitations of using Interferometric Synthetic Aperture Radar(InSAR)in time series analysis is the low-phase coherence associated with rough terrain and vegetated areas,which results in limited spatial coverage in such regions.Permanent scatterers technique was introduced to overcome this limitation using time-series analysis.However,identifying major scatterers within a pixel requires the single-looked pixels oversampling which can be a demanding process especially with large interferometric stacks and vast study areas.Therefore,using multilooked temporal coherent pixels was proposed to increase processing efficiency and coverage by utilizing distributed targets,but this technique may exclude pixels with reliable phase returns because of their temporal varying neighboring pixels.In this paper,we propose a technique to identify multilooked temporal stable pixels with reliable phase returns independent of their neighboring pixels.We conduct a simulation analysis to relate the spatial coherence of a pixel with its expected temporal correlation in the time series analysis module.We found that a liberal temporal correlation threshold of 0.53 in multilooked pixels stack is equivalent to a spatial coherence threshold of 0.2 when using number of looks of 9,which is considered acceptable in temporal coherent pixels,in terms of phase standard deviation.Applying these findings to study the 2011 Tohoku earthquake in the northeastern part of Japan resulted in increasing the number of usable pixels and spatial coverage index by nearly 50.4%and 36.8%,respectively,compared to the temporal coherent pixels.Furthermore,we propose an approach to integrate GPS observations with InSAR time series analysis,which resulted in deformation maps of the megathrust 2011 Tohoku earthquake with mean RMSE of 11.4 mm and a correlation of 98%in comparison to GPS observations.
