Preferencje help
Widoczny [Schowaj] Abstrakt
Liczba wyników

Znaleziono wyników: 38

Liczba wyników na stronie
Pierwsza strona wyników Pięć stron wyników wstecz Poprzednia strona wyników Strona / 2 Następna strona wyników Pięć stron wyników wprzód Ostatnia strona wyników

Wyniki wyszukiwania

help Sortuj według:

help Ogranicz wyniki do:
Pierwsza strona wyników Pięć stron wyników wstecz Poprzednia strona wyników Strona / 2 Następna strona wyników Pięć stron wyników wprzód Ostatnia strona wyników
Airborne laser scanning (ALS) technology allows collecting data describing top forest layer in a very accurate way. This provides a possibility to generate crown height models (CHM) with resolution in which single tree crowns can be detected. In presented study highly automatic algorithm for single tree detection is presented. FALCON II ALS acquired data in Forest Experimental Station in Rogów (central Poland). Data used for algorithm accuracy evaluation were acquired by measuring sample plots on VSD photogrammetric station. CHM with the resolution of 0.5 m has been used in the presented research. Algorithm, in automatic or semi− −automatic way, defines single crowns. Each of the end segments was additionally processed for correctly defined size and area of the crown projection. Received accuracy of correctly detected trees was 69% for all samples (71% for pine stands and 61% mixed stands) in automatic method and 74% (78% and 65% respectively) in semi−automatic method.
Lotniczy skaning laserowy (LIDAR) jest od końca XX w. coraz częściej stosowany w badaniach środowiska naturalnego. Ten aktywny system teledetekcyjny dostarcza bardzo dużej liczby dokładnych danych charakteryzujących badane obiekty oraz udostępnia nowe, dotąd nie eksploatowane płaszczyzny analiz przestrzennych. Prezentowany poniżej tekst jest zbiorem opisów różnego rodzaju metod wykorzystujących dane LIDAR-owe. Mogą być one wykorzystane w szeroko pojętej ochronie środowiska. Nie sposób było wymienić wszystkich zastosowań tego nowego urządzenia. Istotą pracy jest raczej zasygnalizowanie istnienia i możliwości, jakie posiada prezentowana technologia.
Some modern remote sensing technologies, including LIDAR (LIght Detection And Ranging), have significantly developed recently. Laser scanners mounted on the airborne platform make it possible to collect very precise information over large areas, including tree and stand heights. A literature review shows that the model-based method of tree height determination underestimates this parameter in comparison to field measurements. The objective of the study was to analyze accuracy of the automatic height estimation of Scots pine stands, based on the airborne laser scanning data and the example of the Milicz Forest District. Applied algorithm of the stand segmentation into individual trees gave systematic and significant underestimation of the number of trees. The minimum tree height was estimated with a large negative error reaching up to several meters. The maximum mean and top heights were determined more precisely, with a small negative error of a few percent. The sum of tree heights was determined with an error exceeding 40%, which is caused mostly by the error in estimation of the number of trees.
When using such methods as terrestrial laser scanning (TLS), one of the major factors influencing the accuracy of stand characteristics determination is the visibility of trees on a sample plot, which is often obscured by the shadow effect caused by trees located closer to the plot center. Because of this, the percentage of the identified trees and basal area depends on the distance from the plot center: the accuracy of stocking determination decreases as the plot radius increases. The values of such stand characteristics as average breast height diameter, standard deviation of tree diameters and percentiles of the tree diameters' distributions assessed based on all trees and the visible trees only are not significantly different from each other for circular sample plots with 20 m radius. Skewness and kurtosis are not significantly different in plots with radius of 5 and 10 meters. For the 15 m plot radius the difference was significant for about 15% of the analyzed plots. The obtained results correspond with previous findings that report that on the circular sample plots with radius up to 15 m the errors for the number of trees and basal area are relatively small and can be accepted in the practical inventory. The results support the circular sample plots size optimization, including measure− ments performed using a point cloud.
The purpose of this study was to present possibilities of using available Volunteered Geographic Information (VGI) created by users of Flickr to monitor activity in the forest areas within the Warsaw agglomeration. The paper indicates which forest complexes (municipal or agglomeration) were most frequently visited as well as the dates of the greatest use of forest areas in daily, weekly and monthly terms. The study objects include forest areas located in Warsaw and in 52 communes constituting the Warsaw agglomeration. The Kampinos National Park (KPN), which is under strong recreational pressure from the inhabitants of Warsaw agglomeration, was also analysed. In total, we used 1180 images from the Flickr portal in the study. The most visited place was the Kampinos National Park (18.7%), then the forest area in the Legionowo commune (10.7%), which constitute one large forest complex with forests in Choszczówka. Large, compact municipal forest complexes (e.g. Bielański and Linde (8.2%), Sobieski (7.4%) and Kabaty (5.3%)), as well as forests within the Mazowiecki Landscape Park (6.3%) were also very popular. Fraction of photos taken in municipal forests of Warsaw as well as in the agglomeration and KPN forests was larger on Saturdays and Sundays than on business days. It amounted to 51.7% and 59.3%, in forests of a given category respectively. Pictures from the agglomeration and KPN forest areas were most often taken in May (13.8%), while the least often in December (3.3%). Fraction of people visiting municipal forests of Warsaw was the largest in December (12.3%) and the lowest in July (5.0%). On a daily basis, the most activity in both categories was recorded between 11−18. The use of VGI data from the Flickr portal enabled spatial and temporal analysis of user activity in urban and suburban forests. The results obtained confirm current research using survey forms, but in contrast to them, they show the actual places visited for recreation. It should be emphasized that due to the privacy policy of portals, VGI data do not contain information about the metric and status of the user, which makes the analyses inadequate for the entire population.
The paper presents first results of the use of multispectral aerial images to identify the outum phenophases of sessile oak. Observed phenophases are represented with three leaf colors – green, yellow and brown. Color composition of images in three spectral bands: blue, green and red, taken by digital non metric Sigma DP2 cameras, which were carried by Unmanned Aerial Vehicle (UAV) were used. Pictures were taken on 17 October 2011. Two observers made visual crowns classification of 556 oak trees into three groups: green, yellow and brown, on the basis of the dominant color of the leafs. It was found that among observers there is a large compliance in classification (79.7%). Additionally, observations of the spring growth of leafs on 54 trees crowns images recorded from seven positions were evaluated. Although the results may indicate the existence of certain trends, the clear relationship between autumn and spring phases of trees growing can not be noted now (due to small number of sample and short time of observations). The use of UAV to monitor the length of the individual tree growing season has been confirmed.
This paper presents the use of non-metric multispectral digital images analysis acquired by cameras carried by unmanned flying vehicles (UAVs) to assess the density of crowns in Scots pine stands. Images were acquired in October 2011. During the field data acquisition, 272 pine trees were inventoried and classified to 10 classes based on the crown density. These results were compared with the average pixel brightness in four spectral channels B, G, R and IR, collected in three variants – the whole crown, dense part of the crown (the so-called rejection of “outlier” branches) and the central part of the crown with 0.63 m radius. Studies have shown that brightness of images of tree crowns belonging to different classes of density varies insignificantly. This especially concerns the class of trees with a very high degree of defoliation (70–100%) and dead trees. The degree of stabilization can be observed in class 5, which means that a further increase in crown density does not increase the spectral reflectance. Analyzes revealed strong similarity between crowns variants: spectral images obtained for the whole or a part of the crown do not differ from each other. Ultimately, it is difficult to determine which variant is the best for the further classification procedure.
The mainstream of remotely sensed methodology for identifying the tree stand condition is based on spectral responses registered by a multispectral sensor as a digital image. The changes in spectral properties are caused by dying leaves, needles or whole trees. In further steps, the relationship between the spectral values (radiometry) registered in a multispectral satellite image and the health condition of trees should be determined. The most frequent situation includes the one whem dying stand (sensu single tree) occupies the area of <5 m². Therefore the remotely sensed data for determining sanitary conditions of trees must be of a very high spatial resolution (e.g. WorldView2 or 3, GeoEye−1, Pleiades) on one hand and at the same time favourable for the vegetation studies, i.e. utilizing suitable spectral bands and be of low acquisition cost (e.g. RapidEye, LANDSAT−7, ETM +, LANDSAT−8 OLI). Thus a compromise between spatial and spectral resolution should be found to answer the question at what resolution it is possible to clearly separate the damaged tree. The scope of the research included testing of selected methods of satellite image processing and analysis in terms of defining the optimal spatial resolution, which was performed on simulated images obtained for the area of the Beskidy Mountains (S Poland). Pixel size on simulated images was downgraded to the size corresponding to the currently functioning satellite systems. Consequently the obtained material for comparison was free from influence of external factors such as the differences in: time and weather conditions, the geometry of satellite image acquisition, light at the surface of the treetops and phenological vegetation. For each image we used vegetation indices (NDVI and GDVI) and supervised classification. These tests and the obtained results allowed to draw conclusions about the optimal satellite image resolution that can be used to detect damaged or dead stands.
BlackBridge imagery is one of the new means of information used in forest condition analysis. Rapid Eye satellite data with a 5 m spatial resolution register spectral information from 440 to 850 nm through 5 spectral bands. This range of electromagnetic spectrum provides information on plant chlorophyll content as well as cell structure. Such data allows to monitor vegetation condition. This paper focuses on a research conducted in the Sudety and Western Beskidy mountains (southern Poland). The aim of the research was to verify whether high resolution satellite imagery is applicable in detection of the damages caused by Ips typographus and acid rain in Norway spruce dominated stands through supervised classification. BlackBridge Rapid Eye satellite images from 2012 and 2013 were analysed. Various modifications of classification methods were tested, including change in combination of spectral bands. Each method resulted in different classification accuracy. Best results were observed in case of the Maximum Likelihood classification method applied on all spectral bands. The analysis showed that the time of the image registration has a significant impact on classification results. The average classification accuracy for 2012 images was 0.53, whereas for 2013 – 0.69. Moreover, information gathered from 5 m pixels is too general to classify individual dead trees in a precise manner. Tested methods are applicable only in detection of clusters of dead trees.
Pierwsza strona wyników Pięć stron wyników wstecz Poprzednia strona wyników Strona / 2 Następna strona wyników Pięć stron wyników wprzód Ostatnia strona wyników
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.