CC BY
34ENGINEERING SCIENCES
DETAILED MAPPING OF TREE AND SHRUB SPECIES USING HIGH RESOLUTION SATELLITE IMAGERY
Emil Galev
Bulgaria, Sofia, University of Forestry, Assoc. Prof., PhD
Abstract. This paper described the methodology of the vegetation mapping using satellite photo and traditional method combining with remote sensing data. The purpose of the development of the vegetation mapping methodology is to provide an objective and cost-effective survey method utilizing satellite remote sensing for the landscape design projects. IKONOS high-resolution multispectral satellite imagery is used as a technical aid to map ornamental plants. The cardinal study goal at illustrating a concrete instance of applying a new technology in the vegetation inventory. Mapping was carried out using imagery acquired in summer, 2003. The use of satellite photo was fractional, because only crowns with spatial areas greater than 10 square meters were visible on the screen. The result of the analysis shows that image objects extracted from satellite data provide a new opportunity to make detailed inventory maps of ornamental vegetation in parks. Applying the large scale analysis to the images and field surveys, a dendrological map of the study area was obtained. Mapping accuracies were based on ground verified data. Overall, these results confirm that satellite imagery data coupled with field observations and direct measurements can be used effectively for precision mapping of trees and shrubs.
Keywords: detailed mapping, vegetation, inventory, satellite imagery, field surveys.
INTRODUCTION
Detailed mapping of tree and shrub species is often required in various landscape architecture design projects. This paper illustrates a concrete instance of the vegetation mapping using IKONOS data, assuming the future use of satellite imagery data for the vegetation inventory information acquisition. A key requirement for the effective inventory of ornamental plants is the accuracy. Hand-mapping in the field is a technique commonly used in practice, but is inaccurate. In addition hand-mapping from field observation requires access to the site from the ground and visibility, a condition that is not exists always, especially on a very steep ground and in the presence of closely planted trees and bushes. Then measurement of distances is extremely time-intensive and often necessitates of large numbers of linear measurements. Because of these constraints, hand mapping is usually done on an as-needed basis, and comprehensive map that would support data base was done after taking dimensions of the terrain. There is the need, therefore, to develop repeatable and reliable techniques for successive field and computer work, which allows mapping the vegetation objects, which can be accomplished using more conventional methods.
This study tested the suitability of the satellite imagery for improved mapping of plants in the parks. This work will be useful in future decisions about methods for detailed mapping of vegetation.
The objective of the investigation was to make a detailed dendrological inventory. The information it has provided might be useful insofar as directing future surveys or design projects. The graphical output serves as a basic plan to assess future activities in the park. The goals of this project were to map the spatial location of ornamental plants in the study area and to test the efficacy of satellite imagery for plant mapping.
The resolution of IKONOS data is about 1,00 m and provide enough information for distinguishing the majority of the trees (Hiros et al. 2004). On the other hand, IKONOS imagery provides useful information such as diameters of the crowns. Therefore it is suitable for delineating trees in a quite precise manner.
TECHNICAL OBJECTIVES
1. Conduct the inventory in a way that all trees and shrubs occurring at the park are documented with their location and size in a scientifically credible manner.
2. Develop methodology to accurately identifing the location of ornamental park vegetation using satellite imagery, image-editing software and graphic software
3. Test methodology in a study area marked by difficult of access terrain and dense vegetation within the AutoCAD environment.
4. Assess the accuracy of the mapping.
STUDY AREA AND DATA
In this study, a detailed mapping from actual satellite data was conducted for one of the most popular Bulgarian parks. The Architectural and Landscape Gardening Complex "Dvoretza" near the town of Balchik was chosen for the study area. It is a well-known Bulgarian park and is now declared for monument of material culture. The park lies on the Black Sea coast and contains native and Mediterranean ornamental plants, some woodlands and some buildings, pavements, retaining walls and many of architectural elements (all of them are monuments of material culture). This area is characterized by a rich species diversity of trees and shrubs and dense wood massifs. The ground in the area is characterized by too big displacement and terraced configuration.
The remote sensing data used in this study is the IKONOS pansharpen CIR imagery data produced from the panchromatic and multi-spectral IKONOS data taken on July 03, 2003 for the study area. IKONOS data is a commercially available and provides a pixel size 1.00 meter, which makes it suitable for the identification quite a number of vegetation objects and some other objects (Baatz and Schape 2000). In this instance, the IKONOS data was chosen because of its high spatial resolution and profitableness. The higher spectral resolution of the "Quick Bird" data would provide better identification of individual tree crowns, but it is a great expense.
METHODS:
For this study, traditional methods were used in the beginning of the field work for initially mapping of the terrain data, which was recorded in a sketch and then transfered to the digital format. The multispectral satellite data with a spatial resolution of 1.00 m allows identification quite a number of vegetation objects and amend their location or size at a drawing (Usuda 2003).
The sketch-map was created for use as a tentative set. Using a combination of field photos and field notes hand-mapped trees, shrubs and massifs were adjust on the display of the image data. Using distinguished and adjusted trees as pickets were conducted direct measurements in the field and small trees and shrubs were fixed in the drawing.
The final product was derived from the IKONOS data after conducting the field work and after some amendments in the drawing. This product will allow the resource manager of the park to make up to date, well informed management decisions with a detailed vegetation map in a relevant and efficient timeframe.
A detailed drawing for the plants is created in three stages. The investigation started with drawing up a field sketching map produced for the park including areas managed by the park (80 decares). The investigation used standard floristic inventory procedures to identify and document all 274 plant species observed. The locations of the existing plants was judged by sight at first and was recorded as coniferous and deciduous trees or shrubs. The locations was marked in the geodetic survey as points (for the ornamental trees) or as polygons (for woodlands and shrubs). The geodetic survay was updated and made more precise prior to conducting the first stage of field work at the park. The primary purpose of the conducted field inventory is to create a field sketching map, and also verify the current presence of already-documented species, provide information on relative abundance, provide distribution information on endangered or threatened species or species of concern. The place of the plants on the sketching map were fixed by traditional methods, i.e. approximately. The vegetation of the sketch-map has been scanned and processed in a digitized form. The trees were delineated over the drawing like points and the shrubs - like polygons.
IMAGE USING
The study starts with the image conversion in derivative formats for CAD software (Lee et al. 2004), which allow to insert it into AutoCAD where is possible to delineate image objects. The conversion was done using image processing software (Sahin et al. 2004). The satellite photo was not put under geometrical rectification. This was equilibrated with just a little movements of the image in order to coinciding with the geodetic cadastre For this study the satellite photo was used as a technical aid. Its purpose is to provide pickets for the following stage of work - detailed field observations. At this stage, using an electronic ranger were fixed positions of small trees and shrubs toward amended by satellite photo trees. Image insertion implemented in the AutoCAD was used to improve and to specify the field sketching map. Figure 1 shows the results of the intermediate steps of the vegetation inventory map (a fragment of the park area). Figure 1a shows IKONOS image at zoom level equivalent to scale 1:500 on the computer display; Figure 1b illustrates discerning and amendment of some of the trees at the field sketching map, and Figure 1c shows the final vegetation drawing obtained after the last field study.
Fig. 1a. Results of the intermediate steps of the vegetation inventory map (a fragment of the park area). IKONOS image at zoom level equivalent to scale 1:500 on the computer display
Fig. 1b. Discerning and amendment of some of the trees at the field sketching map
Fig. 1c. Vegetation map based on field study
The image was converted to a derivative format for AutoCAD and used to rectification of the location and size of the distinguished tree crowns. Thus the draft of vegetation map was obtained and was finalized and supplemented by subsequent field surveys (field observations and direct measurements). The map compilation took into account massifs and single trees conducted in the area. The method permitted to delineate areas of conifers and deciduous woodlands on an extremely steep and difficult of access terrain, and to identify precise location and size of many individual tree crowns.
Many image objects at this zoom level correspond to individual tree crowns, as well as other spatial objects including roofs of buildings, walks, retaining walls, etc. Some of the shrubs are smaller than the individual pixels. A research question is how to determine an optimal zoom level for identification of individual plants. Many photographs were taken on the terrain and were used as reminders (Figure 2).
Fig. 2. Photographs taken on the terrain just before starting the work. (at a same fragment of the park area)
The purpose of the work was to obtain sound information on the locations and real sizes of the trees and shrubs in a digitized form. The initial objective was to travel all over the entire territory of the park, and to draw up a sketch of the existing vegetation over the geodetic survey.
AMENDMENT IN THE DRAWING
Each graphical object in the hand made sketch needs to be amended, in accordance with the image. Some image objects in this study were relatively easy to be identified based on the spectral properties, and others are difficult to be determined only by the contextual information such as relative sizes, spatial relationships, texture, and so on. The IKONOS data used in this study was taken in July, and vegetation-covered and nonvegetation areas were spectrally distinctive on the imagery. Among the conifer and deciduous trees, and grass areas showed relatively distinctive spectral properties on the photo so that their identification were easier after an arbitrarily variation of their color adjustment. On the contrary, instances of some vegetation classes were difficult to distinguish from other vegetation classes. For instance, some trees (both coniferous and deciduous) with loose crowns are very difficult to identify, but they cast clearly visible shadows on the ground so that that property must be used for the determination rule. For distinguishing deciduous from coniferous trees, texture information of image objects appears to be useful. Figure 1a shows the textural difference between conifer and deciduous trees. Sizes and shapes of image objects were also useful properties for distinguishing some trees such as high specimens with large crowns from others.
In comparison with traditional inventory methods the remote sensing data tend to produce better results representing vegetation cover. In the case of vegetation mapping using very high resolution satellite data, the results are useful in combination with traditional methods results, because the image provides additional information about the location and size of trees in the image objects delineated by the sketch. Anyhow some gaps in the drawing come into sight during the drawing amendment. To rectify these gaps, additional field work is needed, with an electronic telemeter. CONCLUSIONS
Traditional methods for large scale vegetation mapping require expensive, time intensive field surveys. The use of remotely sensed, high resolution, multispectral data for mapping vegetation provides a detailed, accurate product in a time and cost effective manner. For this project, applying a hybrid approach was developed a plant inventory map using imagery data.
The result of the hybrid approach and using IKONOS data in this study suggests that IKONOS data would be a useful additional information source for the vegetation mapping for the landscape architecture design projects.
Very high-resolution satellite images are a useful information source for vegetation mapping, which is part of the design project data set.
Through this project, the spatial distribution of ornamental plants was mapped at a famous Bulgarian park in digital format.
The information collected through this effort will:
• increase the ability of landscape architects to analyze and mapping plants and non vegetative elements of the parks;
• serve as a baseline for long-term monitoring, assist with the characteristics of changes in parks over time and detecting new elements there.
In addition, the data collected through this study will provide the basis for a plant inventory plan for the gardening complex "Dvoretza" near the town of Balchik.
This study shows that ornamental plants in the parks can successfully be mapped using satellite data and conventional methods. These techniques show promise as useful tools for vegetation inventory. This is especially applicable in the landscape architecture design projects because spatial extents and distribution of existing vegetation is very important and define the future plant composition of the parks.
PRODUCTS:
An AutoCAD file including detailed graphical results of the inventory.
An Excel file with a list of plant species documented. The list of plant species documented at the park during the last inventory has been updated. Exactly 274 species are on the list now.
REFERENCES
1. Baatz, M. and Schape, A. (2000) Multiresolution Segmentation - an optimization approach for high quality multi-scale image segmentation. In Angewandte Geographische Informationsverarbeitung, (ed. by Strobl, J., Blaschke, T., and Griesebner, G.), (Heidelberg: Wichmann-Verlag), pp. 12-23.
2. Hirose, Y., Mori, M., Akamatsu Y., Li, Y. (2004) Vegetation Cover Mapping Using Hybrid Analysis of Ikonos Data. Geo-Imagery Bridging Continents XXth ISPRS Congress, 12-23 July 2004 Istanbul, Turkey Commission 7 p. 286 ff.
3. Lee, K.S., Park, Y.I., Kim, S.H., Park, J.H., Woo, C.S., Jang, K.C. (2004) Remote Sensing Estimation of Forest Lai In Close Canopy situation. Geo-Imagery Bridging Continents XXth ISPRS Congress, 12-23 July 2004 Istanbul, Turkey Commission 7 p. 400 ff.
4. Sahin N., Bakici S., Erkek B. (2004) An Investigation on High Resolution Ikonos Satellite Images for Cadastral Applications. Geo-Imagery Bridging Continents XXth ISPRS Congress, 12-23 July 2004 Istanbul, Turkey Commission 7.
5. Usuda, Y. (2003) A study on the optimization of image segmentation in object-Oriented Classification. Proc. annual conference of the Japan Society for Photogrammetry and Remote Sensingin 2003, pp. 125-128.
ПРИМЕНЕНИЕ КОМПЬЮТЕРНЫХ ТЕХНОЛОГИЙ В ЭКОНОМИЧЕСКОМ АНАЛИЗЕ
Абдурахманова Ф. Ф., Нуруллаева Ш. Т., Сайдуллаева С. А.
Узбекистан, Ташкент, Ташкентский государственный экономический университет
Abstract. This article discusses the role of information technology in the economic analysis, the efficiency of analytical work with a computer, the basic requirements for computer analysis.
Keywords: информационные технологии, экономический анализ, персональные электронно-вычислительные машины, методика экономического анализа.
Развитие информационных технологий в Узбекистане является одним из ванных аспектов развития государства и общества. На сегодняшний день невозможно представить деятельность той или иной сферы без применения информационных технологий.
