Time filter

Source Type

Zagreb, Croatia

Klobucar D.,Hrvatske Sume Ltd. | Subasic M.,University of Zagreb | Pernar R.,University of Zagreb
ISPA 2011 - 7th International Symposium on Image and Signal Processing and Analysis | Year: 2011

We present our research on artificial neural network application in remote sensing analysis of forest management data. The presented research is part of our ongoing investigation of texture analysis application on estimation of stand parameters for the forestry needs. In our investigation we have used IKONOS (PAN 1m x 1m) satellite image. We have used two groups of texture features. The first group is based on first and second order histograms and the second group is based on Fourier transform. We have experimented separately with each feature set and also with both of them combined. We tried radial basis neural networks and multilayer perceptrons with different sets of parameters. Optimal network parameters were calculated and we report results of those optimal neural networks. The stand parameters we were estimating include number of trees, stocking, basal area and volume. Each of the parameters is estimated with its own neural network. Separate estimations are done for VI (121 - 140 yrs) and VII (141 - 160 yrs) age class. The experiments have confirmed good estimation accuracy and good correlation with target values. © 2011 University of Zagreb. Source

Klobucar D.,Hrvatske Sume Ltd. | Subasic M.,University of Zagreb
IForest | Year: 2012

A lot of useful data on forest condition can be gathered from the Forest Inventory (FI). Without the help of data analysis tools, human experts cannot manually interpret information in such a large data set. Conventional multivariate statistical analyses provide results that are difficult to interpret and often do not represent the information in a satisfactory way. Our goal is to identify an alternative approach that will enable fast and efficient interpretation and analysis of the FI data. Such interpretation and analysis can be performed automatically with a clustering method, but all clustering methods have some shortcomings. Therefore, our aim was also to provide information in a form suitable for fast and intuitive visualization. Kohonen's Self Organizing Map (SOM) is an alternative approach to data visualization and analysis of large multidimensional data sets. SOM provides different possibilities and our experiments are presented with component matrices of individual stand parameters and label matrices. In forming data clusters, we experimented with hierarchical and non hierarchical clustering methods. Our experiments showed that SOM provides useful information in a form suitable for data clustering and data vi - sualization. This enables an efficient analysis of large FI data sets at different analysis scales. Clustering results obtained with SOM and two clustering al - gorithms are in accordance with ground truth. We have also considered the efficiency of SOM component matrices by visual comparison and correlation among structural parameters and by determining contributions of individual stand parameters to clustering input data. SOM application in visualization and analysis of stand structural parameters enables gathering quickly and efficiently holistic information on the current condition of forest stands and forest ecosystem development. Therefore we recommend the application of Kohonen's SOM for visualization and analysis of FI data. © iForest - Biogeosciences and Forestry. Source

Pandza M.,Murterski skoji Primary School | Milovic M.,Medical and Chemical School | Krpina V.,Hrvatske Sume Ltd.
Natura Croatica | Year: 2011

Floristic researches of the 14 islets and reefs near Vrgada Island (Zadar archipelago, eastern Adriatic) were conducted in the period from 2009 and the spring of 2010. For the 13 islets and reefs, 264 vascular plant taxa were recorded and classified in 184 genera and 63 families. No taxa were recorded for one reef. The domination of therophytes (42.42%) and plants of the Mediterranean floral element (49.62%) confirmed the Mediterranean character of the islets' flora. Source

Degmecic D.,Hrvatske Sume Ltd. | Gros R.,Hrvatske Sume Ltd. | Florijancic T.,Josip Juraj Strossmayer University of Osijek | Ozimec S.,Josip Juraj Strossmayer University of Osijek | Boskovic I.,Josip Juraj Strossmayer University of Osijek
Agriculturae Conspectus Scientificus | Year: 2011

The activity of roe deer was surveyed in five habitats in the Haljevo Forest (Baranja Region, Eastern Croatia), during the 1965/1966 hunting season. The aim was to compare the habitat preferences and to determine differences in the number of animals observed in the study period, by taking into account: period of the year, height of understory layer in forest stands a nd weight of fat deposit around kidney. The animals caught by net were marked and their activities have been observed. A total of 532 sightings of individuals were noted on all five habitats, and the abundance of roe deer was 228 individuals. During the fawning period in the spring, the highest number of animals (n=55) was recorded in black locust stand, followed by oak stands with thick understory layer (n=61) and hornbeam stands with oak (n=51). Regarding the quality of habitat as a food source, the highest number (n=196) was in the hornbeam stand with oak, compared to black locust stand (n=59) and oak stand with thick understory layer (n=20). Fitness of roe deer is estimated by measuring kidney fat from 96 culled individuals. Mean weight of kidney fat was significantly higher in the oak stand with thick understory layer (104.71 g) than in young oak stand (55.83 g). In comparison to black locust stand (81 g) and hornbeam stand with oak (96.46 g), the value was higher but not significantly higher, indicating the importance of the oak acorn in roe deer diet. Source

Discover hidden collaborations