Event

The Department of Food Production at the Faculty of Food and Agriculture will host a lecture on Geometric Features Extraction and Classification of Tree Species. This lecture will be facilitated by Dr. Nicole Christoff ,  Assistant Professor, Technical University of Sofia, Bulgaria. The abstract for the Lecture is share below for your information on friday July 21 at 10 a.m. via zoom

Abstract:

Bulgaria's forests are part of our national wealth and national identity. Their economic, social and environmental functions are important for the sustainable development of society, as well as for the improvement of living conditions, especially in rural and mountain areas. To a large extent, these functions are indispensable not only nationally but also globally.

One of the main problems in the Republic of Bulgaria is the illegal deforestation. Illegal logging in Bulgaria accounts for nearly a third of total production. This affects both the economic losses, which would be around 150 million per year, and the ecosystem. This directly affects the flora and fauna, creating a precondition for deforestation and drying process in coniferous crops, fir and oak forests. This is a prerequisite for erosion, landslides and forest fires, which in turn poses a serious threat to both the condition of forests and human dead’s.

Timber identification methods are divided into manual and automatic. For decades, manual approaches have played an important role in wood identification, such as the magnifying lenses used by anatomists for surface analysis. For species specialists in tropical countries, where there is a great variety of wood, memorizing the microstructure of all species is almost impossible. For example, there are over 1,200 species in India alone. Recognizing large volumes of tree species using the manual approach is sometimes lengthy, time consuming and impossible.

The anatomical characteristics of wood are the most commonly used features used to identify and classify species. Sometimes they are very similar and difficult to distinguish by hand. Automatic approaches use statistical characteristics for analysis and classification. Such an automatic system focuses on statistical analysis of wood pores.

Purpose: To build new theoretical foundations, scientific models, abstractions, as well as rethinking or restoring existing methods and algorithms in order to discover geometric features and classification of tree species.

Achieving this goal involves the following specific tasks: 

• Development of a methodology for taking microscopic images from longitudinal / cross sections of wood samples, based on optical information;

• Creation of a theoretical model delineating the anatomical features common and inherent in tree species; • Development of advanced algorithms for geometric modeling of anatomical features, taking into account the changes of individual tree species and habitats; • Development of effective approaches and methods for detecting anatomical features in order to identify and predict the type of wood.

Meeting Details: