Detection of dead or defoliated spruces using digital aerial data

The purpose of this study was to develop a method for detecting dead and defoliated spruces and defoliated stands in remote-sensing material using a semi-automatic pattern-recognition technique, spectral properties of trees, and different degrees of defoliation. The study material included two mapped defoliation stands in the municipality of Juupajoki (61°50′N, 24°18′E) in southern Finland. The ground truth data were collected during 1996–1997. The aerial color infrared (CIR) photographs, scaled to 1:5000, were taken on 28 June 1995 and on 19 June 1997. The degree of defoliation was visually estimated for every conifer in the defoliation stands. Individual trees in the digital aerial photographs were segmented using a robust segmentation method based on the recognition of tree crown patterns at a sub-pixel accuracy. The images were filtered with a Gaussian N×N smoothing operator, and local maxima above a threshold level were segmented using a directional derivate with some constraints. The segments were placed into defoliation classes using linear Fisher classification models, the parameters of which were estimated by cross-validation. Discriminant analysis was used to find variables for the segment classification. Defoliated tree segments and stands were classified satisfactorily. The accuracy of the pattern-recognition method proved adequate for detecting dead or heavily defoliated trees and heavily defoliated stands. The method described provides an interesting approach to using digital aerial data for automatically detecting severely defoliated spruce stands or individual trees.     

Determination of the spatial distribution of trees from digital aerial photographs

This study examined the possibilities of using computerized digital aerial photograph interpretation in determining the spatial distribution of trees. The material of the study included eight mapped stands in the municipality of Hyytiälä (61°50′N and 24°18′E), in southern Finland. The aerial photographs used were taken in June 1995 at a scale of 1:5000. Two approaches for determining the spatial pattern of trees were used. Firstly, in the point-process based approach used in this study, the individual trees in the digital aerial photograph were segmented by a robust segmentation method, based on recognition of the pattern of tree crowns with sub-pixel accuracy. Secondly, the crown coverage was determined by region growing segmentation combined with active surface representation. The significance of the differences in the means of image coverage pattern indices between the various spatial distribution categories was tested with one-way variance analysis. Because the process misclassified clustered spatial patterns as regular patterns, and regular patterns as random patterns, the usability of digital aerial photographs seems to be limited for the point-process based determination of the spatial pattern of trees if the scale is 1:5000 or less. When image coverage pattern indices were applied, the differences in the means of the spatial distribution categories proved not to be clearly statistically significant due to the great variation within classes. However, interpretation of crown coverage could have applications in practical forestry due to the low resolution requirements for the images used.     

Aerial photo-interpretation using Z/I DMC images for estimation of forest variables

Abstract

Interpretation and tree height measurements in aerial photographs using photogrammetric workstations are frequently performed in standwise forest inventory. Images acquired by digital aerial cameras are now replacing the traditional film-based aerial photographs. In this study, digital images from the airborne Z/I DMC system for standwise estimation of stem volume, tree height and tree species composition were investigated at a 1200 ha forest area located in southern Sweden (58°30′N, 13°40′E). The 56 selected stands were dominated by Norway spruce [Picea abies (L.) Karst.] and Scots pine (Pinus sylvestris L.) with stem volume in the range of 30–630 m³ ha^?1 (average 300 m³ ha^?1) and tree height in the range of 6–28 m (average 20 m). The large-format pansharpened colour infrared images were captured at a flight altitude of 4800 m above ground level corresponding to a pixel size of 0.48 m. The photo-interpretation was conducted by four professional interpreters, independently. In particular, two different base-to-height ratios (i.e. the ratio between the ground distance between image centres at the time of exposure and the flight altitude above ground level) of 0.26 and 0.39 were evaluated, but no significant difference in the estimation accuracy for stem volume and tree height was found. The accuracy for stem volume estimation in terms of relative root mean square error, corrected for systematic errors, was on average 24% (in the range of 17–39%). The corresponding accuracy for tree height estimation was on average 1.4 m (in the range of 0.9–1.6 m). The tree species composition accuracy assessment using a fuzzy set evaluation procedure showed that 95% of the stands were correctly classified. The estimation accuracies are in agreement with previous results using conventional film-based aerial panchromatic photographs.     

Forest variable estimation using photogrammetric matching of digital aerial images in combination with a high-resolution DEM

Abstract

The rapid development in aerial digital cameras in combination with the increased availability of high-resolution Digital Elevation Models (DEMs) provides a renaissance for photogrammetry in forest management planning. Tree height, stem volume, and basal area were estimated for forest stands using canopy height, density, and texture metrics derived from photogrammetric matching of digital aerial images and a high-resolution DEM. The study was conducted at a coniferous hemi-boreal site in southern Sweden. Three different data-sets of digital aerial images were used to test the effects of flight altitude and stereo overlap on an area-based estimation of forest variables. Metrics were calculated for 344 field plots (10 m radius) from point cloud data and used in regression analysis. Stand level accuracy was evaluated using leave-one-out cross validation of 24 stands. For these stands the tree height ranged from 4.8 to 26.9 m (17.8 m mean), stem volume 13.3 to 455 m³ ha^?1 (250 m³ ha^?1 mean), and basal area from 4.1 to 42.9 m² ha^?1 (27.1 m² ha^?1 mean) with mean stand size of 2.8 ha. The results showed small differences in estimation accuracy of forest variables between the data-sets. The data-set of digital aerial images corresponding to the standard acquisition of the Swedish National Land Survey (Lantmäteriet), showed Root Mean Square Errors (in percent of the surveyed stand mean) of 8.8% for tree height, 13.1% for stem volume and 14.9% for basal area. The results imply that photogrammetric matching of digital aerial images has significant potential for operational use in forestry.     

Predicting species-specific basal areas in urban forests using airborne laser scanning and existing stand register data

The aim of this work was to examine how well species-specific stand attributes can be predicted using a combination of airborne laser scanning (ALS) and existing stand register data in urban forests. In this context, the ability of three data combinations: ALS data and stand register data, ALS data and digital aerial images and all of these combined, was tested in the prediction of species-specific basal areas. We divided tree species into seven and three different tree species strata and applied two prediction methods: (1) regression method, in which the predicted total basal area was divided into tree species based on tree species proportions from stand register data, and (2) the nearest neighbour (NN) method, in which tree species proportions were used as predictor variables for species-specific basal areas. Prediction models were built based on training data of 205 field plots, and the accuracy of the models was tested based on validation data of 52 forests stands. Our results showed that species-specific predictions of seven tree species were more accurate when tree species proportions from stand register data were used in the prediction. Both the regression and the NN method provided reasonable accuracy. This study showed that tree species information from existing stand register data could be used as an alternative for aerial images in ALS-based forests inventories. The use of ALS data together with stand register data and small field data could also be economically beneficial in an inventory of urban forests.     

Determination of Mean Tree Height of Forest Stands by Digital Photogrammetry

The mean tree height of 73 forest stands in a 1000 ha forest area was determined from canopy heights generated by automatic image matching using a digital photogrammetric workstation and digitized panchromatic aerial photographs with a scale of 1:15 000. First, the mean height of each stand was computed as the arithmetic mean of the quantile corresponding to the 75th percentile of the distribution of the canopy heights from the image matching within square grid cells with cell sizes of 236-400 m². The mean heights from the image matching underestimated the true heights by 5.42 m. Secondly, field-measured mean tree heights of 165 georeferenced sample plots distributed systematically throughout the 1000 ha forest area were regressed against the mean heights derived from the image matching. The regression equations were used to predict the mean heights of the 73 stands. In very young forest stands, the predicted mean heights overestimated the true heights by 0.4 m and the precision was 0.9-1.0 m. In young and mature stands, the average difference between predicted height and ground-truth ranged between -1.6 and 0.5 m, and the precision ranged from 1.1 to 2.1 m.     

Stem Diameter Estimation from Aerial Photographs

Remote sensing techniques have proven successful for producing stem maps of forests in leaf-on condition from high-resolution imagery. This paper demonstrates how a mathematical model for the surface of a stem can be used to estimate the breast-height diameter of individual trees from aerial photographs to give information on basal area. The diameters are estimated by likelihood estimation from images of a forest in leaf-off condition where the stems and their shadows are visible. Applied to a homogeneous and monospecific oak ( Quercus robur L.) stand under standard silvicultural treatment in Denmark, the estimation was successful for 56 out of 60 trees. The root mean squared error on the diameter was 4.2 and 3.2 cm using three and five images, respectively. The key conclusion is that it is feasible to infer fairly accurate information about the diameters and three-dimensional positions of stems from aerial photographs.     

Using pre-classification to improve the accuracy of species-specific forest attribute estimates from airborne laser scanner data and aerial images

The aim of this study was to examine whether pre-classification (stratification) of training data according to main tree species and stand development stage could improve the accuracy of species-specific forest attribute estimates compared to estimates without stratification using k-nearest neighbors (k-NN) imputations. The study included training data of 509 training plots and 80 validation plots from a conifer forest area in southeastern Norway. The results showed that stratification carried out by interpretation of aerial images did not improve the accuracy of the species-specific estimates due to stratification errors. The training data can of course be correctly stratified using field observations, but in the application phase the stratification entirely relies on auxiliary information with complete coverage over the entire area of interest which cannot be corrected. We therefore tried to improve the stratification using canopy height information from airborne laser scanning to discriminate between young and mature stands. The results showed that this approach slightly improved the accuracy of the k-NN predictions, especially for the main tree species (2.6% for spruce volume). Furthermore, if metrics from aerial images were used to discriminate between pine and spruce dominance in the mature plots, the accuracy of volume of pine was improved by 73.2% in pine-dominated stands while for spruce an adverse effect of 12.6% was observed.     

基于多特征提取与注意力机制深度学习的高分辨率影像松材线虫病树识别

为了提高松材线虫病树的监测效率,减少其对林业生产造成的损失,提出一种基于多特征提取与注意力机制深度学习的高分辨率影像松材线虫病树识别方法.该方法首先在高分辨率遥感影像上提取松材线虫病树的光谱特征、空间特征等多特征,然后进行Relief特征选择算法,取特征权重前8个特征进行病树识别,发现选择差值植被指数DVI(diffe...     

An aerial photograph auto-interpretation system for forest resource inventory

Aerial photographs arc important information sources for forest inventory. With the development of science and technology, an automated approach to aerial photograph interpretation has been sought. This paper introduces an experimental computer system for the auto-interpretation of aerial photographs. On thebasis of the experimental system, several algorithms are developed for the interpretation of density, crown closure, working groups and species composition of forest stands. Experiments show that auto—interpretation works well for some forest inventory data.     

Deciduous forest mapping using change detection of multi-temporal canopy height models from aerial images acquired at leaf-on and leaf-off conditions

Discrimination of deciduous trees using spectral information from aerial images has only been partly successfully due to the complexity of the reflectance at different view angles, times of acquisition, phenology of the trees and inter-tree radiance. Therefore, the objective was to evaluate the accuracy of estimating the proportion of deciduous stem volume (P) utilizing change detection between canopy height models (CHMs) generated by digital photogrammetry from leaf-on and leaf-off aerial images instead of using spectral information. The study was conducted at a hemi-boreal study area in Sweden. Using aerial images from three seasons, CHMs with a resolution of approximately 0.5?m were generated using semi-global matching. For training plots, metrics describing the change between leaf-on and leaf-off conditions were calculated and used to model the continuous variable P, using the Random Forest approach. Validated at sub-stands, the estimation accuracy of P in terms of root mean square error and bias was found to be 18% and ?6%, respectively. The overall classification accuracy, using four equally wide classes, was 83% with a kappa value of 0.68. The validation plots in classes of high proportion of coniferous or deciduous stem volume were well classified, whereas the mixed forest classes showed lower classification accuracies.     

Estimating forest characteristics in scanned aerial photographs with respect to requirements for economic forest management planning

The objective of forest management planning is often expressed as maximum sustainable economic yield. Methods used to collect information for forestry planning should, therefore, include variables significant for economic evaluations of management alternatives. It is important to be able to differentiate mature stands with respect to timber volumes and species mixture. In this study, digital high‐altitude aerial photographs are tested as a data source for planning. Circular plot data from a forest estate in northern Sweden were used as reference material. Global positioning system (GPS) measurements, with differential correction, were used to georeference the plots. Harvesting priorities were calculated for each plot using the Forest Management Planning Package. Volumes, species mixture and harvest priorities were estimated using regression analysis based on textural and spectral information from aerial photographs. The results show that the dependent variables could be estimated fairly well using only spectral information, e.g., R ² = 0.44 when estimating timber volume at reference plot (10 m radius) level. Aggregated to stand level, the precision was comparable with customary field survey methods (e.g., RMSE= 13.4% for timber volume).     

Mapping site index and age by linking a time series of canopy height models with growth curves

Site index (SI) is one of the main measures of forest productivity in North America. For monospecific even-age stands, it is defined as the height of dominant trees at a given reference age or presented as an age–height curve. SI normally reflects the overall effect of all the environmental parameters that determine height growth locally. However, measuring SI can only be achieved though field observations and is, for this reason, limited to sample plots. In this study, we propose a new method for quantifying and mapping SI and age based on known age–height curves and time series of canopy height models (CHMs) produced using digital photogrammetry and lidar. Digital surface models (DSMs) are created by applying an automated stereo-matching algorithm to scanned aerial photographs. The canopy height is obtained by subtracting the lidar ground elevations from the DSM. Using aerial photographs covering the 1945–2003 interval and a recent lidar coverage, CHMs could be reconstructed retrospectively for a period of over 58 years. Regionally calibrated age–height curves were fitted to observations that were extracted cell-wise from the historical CHMs to estimate SI and age values for all undisturbed locations. Results demonstrate that SI and age of jack pine (Pinus banksiana [Lamb.]) stands can be quantified respectively with an average bias of 0.76 m (2.41 m root mean squared error, RMSE) and 1.86 years (7 years RMSE). The method can be used to produce quasi-continuous maps of SI and age and to estimate productivity in a spatially explicit way.     

Comparison of tree growth and undergrowth development in aerially seeded and planted Pinus tabulaeformis forests

Direct seeding is a less expensive practice than planting and has the potential to become a viable alternative to transplanting for afforestation and regeneration purposes. As an effective and a less costly regeneration method, aerial seeding has been applied with several tree species. As early as 1956, Chinese people engaged in aerial seeding and stands with a total of 2.97×10⁷ hm² have been developed up to 2004. Our study tested whether the growth of planted Chinese pine (Pinus tabulaeformis Carr.) seedlings and its undergrowth development in northwest aspects differ from that of aerially sown seedlings on the northern and northwestern aspects of slopes. In 2007, we collected data such as height, diameter at breast height (DBH), clear bole height and canopy widths of trees, abundance, coverage, and frequency of shrubs and herbs from 21-year-old planted Chinese pine stands on a northwestern aspect (PNW), aerially sown stands in a northwest aspect (ANW) and aerially sown stands in a northern aspect (AN). Results showed that the relation of crown area and mean DBH was best fitted by a double inverse model for the ANW and AN forests and by a quadratic model for the PNW forest. There was no difference in the growth between ANW and AN forests, while growth was significantly higher in the PNW forest than in the ANW and AN forests. That was consistent with the Sorenson diversity indices in the shrub and herb layers, indicating that there was a large number of the same species in both aerially seeded stands, although their locations were different. Both the number of species in the undergrowth and the Shannon-Wiener index in the shrub layer were higher in the PNW stands than in the ANW and AN stands. Dominant families for all three stands were Rosaceae and Compositae in the shrub and herb layer, respectively. The dominant species for all three stands was Spiraea pubescens in the shrub layer, while the dominant species was different from each other in the three stands. The discrepancy in diversity and composition of species in the herb layer show that herbs are sensitive to shrubs in the three forests. High mortality and skewed diameter distributions reflect severe competition and too high a density in the aerially seeded forests. Thus, aerial seeding is a viable and effective regeneration technique, but management practices, such as thinning, should be applied to these forests.     

A digital photographic method for 3D reconstruction of standing tree shape

A digital photographic method is presented which is able to reconstruct the profile of the stem on standing trees up to a height of 12 m and to provide a fine level of detail. The method uses two digital photographs taken at 90° to each other and does not require special illumination conditions. A method is proposed to perform the data acquisition process from the two photographs and to transform the stem dimensions and 3-D position from pixels to units of length. The accuracy of this method for measuring tree shape was tested by comparing the results with those obtained from a laser system. The comparison showed that the photographic method provides a good assessment of standing tree shape.     

Comparing biophysical forest characteristics estimated from photogrammetric matching of aerial images and airborne laser scanning data

Recent development in aerial digital cameras and software facilitate the photogrammetric point cloud as a new data source in forest management planning. A total of 151 field training plots were distributed systematically within three predefined strata in a 852.6 ha study area located in the boreal forest in southeastern Norway. Stratum-specific regression models were fitted for six studied biophysical forest characteristics. The explanatory variables were various canopy height and canopy density metrics derived by means of photogrammetric matching of aerial images and small-footprint laser scanning. The ground sampling distance was 17 cm for the images and the airborne laser scanning (ALS) pulse density was 7.4 points m^–2. Resampled images were assessed to mimic acquisitions at higher flying altitudes. The digital terrain model derived from the ALS data was used to represent the ground surface. The results were evaluated using 63 independent test stands. When estimating height in young forest and mature forest on poor sites, the root mean square error (RMSE) values were slightly better using data from image matching compared to ALS. However, for all other combinations of biophysical forest characteristics and strata, better results were obtained using ALS data. In general, the best results were found using the highest image resolution.     

Derivation of stumpage value and logging costs for individual forest stands using aerial photographs and cartographic modelling

A procedure for calculation of stumpage value and logging costs of individual mature forest stands using aerial photo‐interpretation and a grid‐based geographical information system (GIS) is presented. The stumpage value may be computed from characteristics related to the trees, i.e. site index, stand mean height, crown closure, and tree species distribution, using price equations. By means of cost equations, the logging costs may be calculated from the tree characteristics and the terrain characteristics of slope gradient and skidding distance. The practical application of the procedure was demonstrated by a case study in a 710 ha forest area in southern Norway. The tree characteristics were determined by photo‐interpretation of individual stands. Skid paths for wood transportation from the stands to landings along the forest roads were delineated by photo‐interpretation of the ground conditions. Slope and skidding distances were derived by a digital elevation model and cartographic modelling. Finally, the photo‐interpreted tree characteristics and the computed slope and skidding distances were used for calculation of the stumpage value and the logging costs of each stand. According to previous tests, the accuracy of the procedure corresponded to the accuracy that could be achieved by the field‐survey methods used most frequently.     

The Ingrowth Bag Method in Measuring Root Production on Peatland Sites

The aim of this study was to develop a method for segment-based forest inventory and determine whether segment-level inventories can be used in forest management planning. The study area covered 76 ha located in two different aerial photographs in eastern Finland. The study area was segmented into 220 segments with the aid of aerial photographs and the segment-level forest characteristics were assessed in the field using relascope sample plots and a field computer which displayed the aerial photographs, segment borders and surveyor's location on the screen. The segment estimates were calculated as weighted averages of k nearest neighbours (kNN) for the segments and the sample plots. The estimates were tested with a cross-validation technique. The averages and the standard deviations of the spectral values of aerial images extracted for the segments and the sample plots were used in the kNN estimation. The relative root mean square error of the mean volume was 58.1% (bias –6.4%) at the segment level and 57.9% (bias –0.9%) at the sample plot level. The segment-based approach studied here needs further research and improvement before it can be applied to forest management planning.     

Stand structure and successional trends in virgin boreal forest reserves in Sweden

Fire history and stand structure was examined in twelve virgin forest stands situated within forest reserves in northern Sweden. The selected stands represented fire refuges as well as different successional stages after fire. Six of the stands were dominated by Norway spruce (Picea abies L. Karst.), three were dominated by Scots pine (Pinus sylvestris L.), and three were dominated by hairy birch (Betula pubescens Ehrh.) or aspen (Populus tremula L.). In 3 of the southernmost stands, the average fire interval was 34 to 65 years during the late 1600s to late 1800s, but since 1888 no fires had occurred in any of the stands. The absence of fire disturbance since 1888 is probably caused by the fire suppression in the overall landscape. The standing volume of living trees ranged between 87 and 511 m³ ha⁻¹ while the volume of dead trees, including both snags and logs, ranged between 27 and 201 m³ ha⁻¹. The volume of dead trees constituted ca. 30% of the total stem volume. In the conifer dominated stands, there was a statistically significant relationship between total stem volume, including both living and dead trees, and site productivity. A comparison between the amount of dead and living trees indicated substantial changes in tree species composition in several stands. It is suggested that data on the amount of dead trees, especially logs, and its distribution over decay classes could be used to examine the continuity of certain tree species. All stands had a multi-sized tree diameter distribution, which in most cases was similar to a reversed J-shaped distribution. In general spruce was numerous in the seedling cohort and in small diameter classes, indicating that its proportion in the stands was stable, or was increasing at the expense of pioneer tree species such as pine, aspen and silver birch (Betula pendula Roth.). The most numerous species in the seedling cohort, rowan (Sorbus aucuparia L.), was almost totally missing in the tree layer, indicating a high browsing pressure preventing rowan seedlings from growing into trees. The general increase of spruce and the sparse regeneration of pioneer species, in the stands previously affected by fire, are discussed in relation to natural disturbance regimes, biological diversity and nature conservation policies. It is proposed that reintroduction of fire disturbance is a necessity for future management plans of forest reserves. Other management practices to increase species diversity within forest reserves are also discussed.