Aristotle University Of Thessaloniki (AUTH)
Εκτίμηση της βιομάζας δασικών συστάδων με τη χρήση μεθόδων τηλεπισκόπησης και γεωγραφικών συστημάτων πληροφοριών
Abstract
dc:descriptionThe importance of forests, in both economic and environmental point of view, has made imperative the need of change of forestry policy. The implementation of ecosystem management, shifted the emphasis from maintaining the ability to harvest, to the maintenance of healthy and diverse ecosystems. In this frame, the biomass as a measure of production, assembles more advantages, in relation with the traditional measurements of volume. The traditional methods for biomass estimation, based directly or indirectly to the field measurements, are the more accurate techniques for the collection of biomass data. But they present some inherent disadvantages concerning mainly the high cost for a large scale application. The remote sensing based methods, could be an alternative source of information, for the estimation of quantity and distribution of forest biomass. The way with which the forest vegetation is captured from a remote sensing image, is a result of species composition, development stage, structure, etc. Within this sense, the objectives of this thesis, was firstly to evaluate the general ability of remote sensing methods to estimate forest biomass. Also attempts, to examine the degree in which different remote sensing data and various techniques of analysis can be performed in the field of forest biomass estimation. From the results proved that remote sensing has the ability to estimate forest biomass in a significant level. Spectral information of the thematic mapper of Landsat, as well as spatial information of IKONOS images, show that, with the selection of the appropriate techniques, in both pre-processing and processing stages, can predict biomass variability with a success of more than 60%. For the Landsat multispectral image, the most valuable source of information proved to be the middle infrared (MIR) band (TM5). MIR has the ability to reveal information about the fraction of broadleaved understory vegetation that participate in the spectral response of the forest stand. The fraction of understory vegetation in the field of view of the image, is necessary information for the biomass estimation for a coniferous stand. Best predictor from a series of vegetation indices proved to be the RDVIc which is a combination of the classic index RDVI, weighted with the quantity of MIR reflected from the forested areas. Interesting results was also revealed from the spatial analysis, extracting texture measures from panchromatic IKONOS image. The measure of Homogeneity estimated that can predict biomass variability with success of 60%. Homogeneity probably related with the structure of overstory vegetation. Regression model based mainly in the presence of homogeneity as independed variable predicted biomass with precision of 82%. Finally it has to be pointed out the importance in the selection of an appropriate topographic correction method. Topographic correction method, proved that has a direct relation with the improvement of the quality of image spectral information. Methods that takes into account the behaviour of reflected radiation due to the roughness of surface, as the Minnaert method, can improve the accuracy of the results more than 50%, in relation with method which based only to the geometry of the surface.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2007
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Makras, Apostolos
- Μάκρας, Απόστολος
Subjects
dc:subject × 17- Εκτίμηση δασικής βιομάζας
- Δάση κωνοφόρων
- Θεματικός χαρτογράφος (ΤΜ)
- Προεπεξεργασία εικόνων
- Δείκτες βλάστησης
- Ανάλυση υφής
- Forest biomass estimation
- Conifer forests
- Thematic mapper
- Image preprocessing
- Vegetation indices
- IKONOS
- Texture analysis
- Γεωπονικές Επιστήμες και Κτηνιατρική
- Γεωπονία, Δασολογία και Αλιεία
- Agricultural and Veterinary Sciences
- Agriculture, Forestry and Fisheries
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/19200