| 1. |
- Granholm, Ann-Helen, et al.
(författare)
-
Estimating vertical canopy cover using dense image-based point cloud data in four vegetation types in southern Sweden
- 2017
-
Ingår i: International Journal of Remote Sensing. - : Informa UK Limited. - 0143-1161 .- 1366-5901. ; 38, s. 1820-1838
-
Tidskriftsartikel (refereegranskat)abstract
- This study had the aim of investigating the utility of image-based point cloud data for estimation of vertical canopy cover (VCC). An accurate measure of VCC based on photogrammetric matching of aerial images would aid in vegetation mapping, especially in areas where aerial imagery is acquired regularly. The test area is located in southern Sweden and was divided into four vegetation types with sparse to dense tree cover: unmanaged coniferous forest; pasture areas with deciduous tree cover; wetland; and managed coniferous forest. Aerial imagery with a ground sample distance of 0.24 m was photogrammetrically matched to produce dense image-based point cloud data. Two different image matching software solutions were used and compared: MATCH-T DSM by Trimble and SURE by nFrames. The image-based point clouds were normalized using a digital terrain model derived from airborne laser scanner (ALS) data. The canopy cover metric vegetation ratio was derived from the image-based point clouds, as well as from raster-based canopy height models (CHMs) derived from the point clouds. Regression analysis was applied with vegetation ratio derived from near nadir ALS data as the dependent variable and metrics derived from image-based point cloud data as the independent variables. Among the different vegetation types, vegetation ratio derived from the image-based point cloud data generated by using MATCH-T resulted in relative root mean square errors (rRMSE) of VCC ranging from 6.1% to 29.3%. Vegetation ratio based on point clouds from SURE resulted in rRMSEs ranging from 7.3% to 37.9%. Use of the vegetation ratio based on CHMs generated from the image-based point clouds resulted in similar, yet slightly higher values of rRMSE.
|
|
| 2. |
- Granholm, Ann-Helen, et al.
(författare)
-
Estimating vertical canopy cover with dense point cloud data from matching of digital aerial photos
- 2015
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- This study aims to explore the use of dense point clouds from matching of aerial photos for estimation of vertical canopy cover (VCC), defined as the proportion of the forest floor covered by the vertical projection of the tree crowns. VCC is commonly estimated using vegetation ratio (VR) derived from airborne laser scanner (ALS) data. A reliable measure of VCC from matching aerial photos would aid in vegetation mapping and reduce the need for repeated ALS data acquisition. The test area is located in southern Sweden and covers a variety of vegetation types. In total 367 sample plots were placed in parts of the study area representing VCC ranging from 0 % up to close to 100 %. ALS data with a density of 20 returns per m2 was used for calculating the VR as the proportion of first returns above a threshold. Aerial imagery with a ground sample distance of 0.25 m was matched to produce dense point cloud data, which was used to derive digital surface models (DSMs) with grid size from 0.25 m up to 2.0 m. Local maxima (LM) detection was applied to the DSMs with search windows of 0.5 m size up to 2.0 m. The heights of the LM were normalized using a digital elevation model (DEM) derived from ALS data. Regression analysis was applied with the VR as dependent variable and the sum of the height of LM within sample plots as independent variable. Results from linear regression using heights of LM detected in a DSM of 0.25 m resolution with a 0.5 m search window gave an root mean square error (RMSE) of 5.5 % and relative RMSE (rRMSE) of 9.3 % in forest on rocky outcrops and boulders, while wooded pasture gave RMSE = 6.3 % and rRMSE = 19 %.
|
|
| 3. |
- Granholm, Ann-Helen
(författare)
-
Segmentation of forest patches and estimation of canopy cover using 3D information from stereo photogrammetry
- 2016
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- 3D information extracted by image matching of aerial images, so called image-based point clouds, have been found to provide accurate vegetation height measurements. This has led to an increased interest from the vegetation mapping community, since aerial images are an affordable alternative to airborne laser scanner (ALS) data. In Sweden, this is especially interesting due to the National Mapping Agency’s decision to derive 3D information from annually acquired aerial imagery, starting in 2016. Previous studies have shown that image-based point cloud data derived from standard stereo aerial images is of potential use for forest inventory and change detection. In this thesis, the focus is on exploring the utility of image-based point clouds, and surface models, for vegetation mapping; more specifically, it explores segmentation of vegetation patches based on height above ground, estimation of tree height, and estimation of vertical canopy cover. The studies were conducted in a study area located in the hemi-boreal zone of southern Sweden. Segmentation based on canopy height models (CHMs) derived by image matching combined with a digital elevation model (DEM) from ALS data was found to deliver polygons within which tree height varied with a few meters. Tree height was estimated using height percentiles derived from the CHM and the results were similar to previous studies using image-based point clouds. Estimation of vertical canopy cover resulted in low accuracy due to underestimation when the canopy cover was sparse, and overestimation when the canopy cover was dense, while behaving linearly at approximately 15 – 85 % canopy cover. Dominant tree species influenced the results of estimation of tree height, as well as vertical canopy cover. Vegetation mapping using image-based point cloud data holds great potential and further research is needed to gain knowledge of appropriate methods and limitations.
|
|
| 4. |
- Granholm, Ann-Helen, et al.
(författare)
-
The potential of digital surface models based on aerial images for automated vegetation mapping
- 2015
-
Ingår i: International Journal of Remote Sensing. - : Informa UK Limited. - 0143-1161 .- 1366-5901. ; 36, s. 1855-1870
-
Tidskriftsartikel (refereegranskat)abstract
- Segmentation of vegetation patches was tested using canopy height models (CHMs) representing the height difference between digital surface models (DSMs), generated by matching digital aerial images from the Z/I Digital Mapping Camera, and a digital elevation model (DEM) based on airborne laser scanner data. Three different combinations of aerial images were used in the production of the CHMs to test the effect of flight altitude and stereo overlap on segmentation accuracy. Segmentation results were evaluated using the standard deviation of photo-interpreted tree height within segments, as well as by visual comparison to existing maps. In addition, height percentiles extracted from the CHMs were used to estimate tree heights. Tree height estimation at the segment level yielded root mean square error (RMSE) values of 2.0 m, or 15.1%, and an adjusted coefficient of determination (adjusted R2) of 0.94 when using a CHM from images acquired at an altitude of 1200 m above ground level (agl) and with an along-track stereo overlap of 80%. When a CHM based on images acquired at 4800 m agl and an overlap of 60% was used, the corresponding results were an RMSE of 2.2 m, or 16.0%, and an adjusted R2 of 0.92. Tree height estimation at the plot level was most accurate for densely forested plots dominated by coniferous tree species (RMSE of 2.1 m, or 9.8%, and adjusted R2 of 0.88). It is shown that CHMs based on aerial images acquired at 4800 m agl and with 60% along-track stereo overlap are useful for the segmentation of vegetation and are at least as good as those based on aerial images collected at a lower flight altitude or with greater overlap.
|
|
| 5. |
- Nilsson, Mats, et al.
(författare)
-
Computer classification of General Habitat Categories by combining LiDAR and SPOT data
- 2010
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- General Habitat Categories (GHC) is a classification scheme developed in BioHab1, 2 and a central concept in EBONE3. A characteristic of GHC is plant height, which can be derived using Light Detection And Ranging (LiDAR) data. Computer classification of GHCs might be improved by combining spectral information in optical satellite data with LiDAR. The aim of this pilot study was to investigate to which degree airborne LiDAR improves SPOT data based classification of a selection of GHCs in a forest area in southern Sweden. Lat. 58° 30’ N Long 13° 40’ E. Managed forest with Scots pine (Pinus Sylvestris), Norway spruce (Picea Abies) and birch (Betula spp). A SPOT 5 HRG XS scene. Airborne LiDAR data with an average point density of 26 returns/m2. Photo interpretation of GHCs, 585 sample plots, in aerial DMC images. Combining LiDAR and SPOT data shows promise, considering the restrictions to this study. In a similar study, using the same dataset for classifying CORINE land cover types, overall accuracy increased from 67.1% to 77.6% when adding LiDAR data4. This means that there is potential, though the methods need improvement and further tests should include a larger test area providing adequate amounts of sample plots per GHC.
|
|
| 6. |
- Nordkvist, Karin, et al.
(författare)
-
Classification of EBONE General Habitat Categories in a Swedish forest area by using LiDAR in combination with SPOT satellite data
- 2010
-
Rapport (övrigt vetenskapligt/konstnärligt)abstract
- We have investigated to which degree a combination of optical satellite data and LiDAR data can improve classification accuracy of the General Habitat Categories (GHC) used by the FP7 project European Biodiversity Observation Network (EBONE), compared to using satellite data alone. The study was carried out in Remningstorp, a forest dominated area in southern Sweden. Remote sensing data used in the study were a SPOT 5 image from August 2009 and a laser scanning (26 points/m2) from September 2008. Ground truth samples were collected by interpretation of color infrared digital air photos from September 2009. Maximum likelihood and Random Forests classifications were made with satellite data and with a combination of satellite and LiDAR data. The classification scheme consisted of six forest classes, arable land and pasture land. The use of LiDAR data improved over-all accuracy with 6% for maximum likelihood classification and 7% for Random Forests. The highest over-all accuracy was obtained with Random Forests, but on the expense of the smaller classes.
|
|
| 7. |
- Nordkvist, Karin, et al.
(författare)
-
Combining optical satellite data and airborne laser scanner data for vegetation classification
- 2012
-
Ingår i: Remote Sensing Letters. - : Informa UK Limited. - 2150-704X .- 2150-7058. ; 3, s. 393-401
-
Tidskriftsartikel (refereegranskat)abstract
- The aim of this study was to investigate to which degree the accuracy of vegetation classification could be improved by combining optical satellite data and airborne laser scanner (ALS) data, compared with using satellite data only. A Satellite Pour l'Observation de la Terre (SPOT) 5 scene and Leica ALS 50-II data from 2009, covering a test area in the mid-Sweden (latitude 60 degrees 43' N, longitude 16 degrees 43' E), were used in maximum likelihood and decision tree classifications. Training and validation data were obtained from the interpretation of digital aerial photo stereo models. Combination of SPOT and ALS data gave classification accuracies up to 72%, compared with 56% using only SPOT data. This indicates that integrating features from large area laser scanning may lead to significant improvements in satellite data-based vegetation classifications.
|
|
| 8. |
|
|
