We found a match
Your institution may have rights to this item. Sign in to continue.
- Title
Wood Types Classification using Back-Propagation Neural Network based on Genetic Algorithm with Gray Level Co-occurrence Matrix for Features Extraction.
- Authors
Santosa, Stefanus; Pramunendar, R. A.; Prabowo, D. P.; Santosa, Yonathan P.
- Abstract
Tree species are very diverse with more than 15,000 species are known, and mostly found in rain forest such as Indonesian rain forest. Species with similar morphological and anatomical features are very difficult to be classified manually. Manual classification for wood production is time and resources consuming, and the result is less accurate. Wood classification have significant effect on the processing and the quality of industrial wood products. Research on the automation of wood type identification has been done but still not able to cover various the type of wood from different tree species. This research is focused on creating a model for identification of some rain forest trees such as teak (Tectona grandis), albizia (Albizia chinensis), mahogany (swietenia mahagoni), and melia (Melia azedarach). Classification was performed using back-propagation neural network with genetic algorithm optimization and gray level co-occurrence matrix for feature extraction. Genetic algorithm was used to select which GLCM features function should be used as neural network inputs. IDM, correlation and entropy was chosen the most and by doing so, we are able to achieve higher accuracy compared to the previous model. Furthermore, deeper analysis on the angle of GLCM in wood classification shows that the best accuracy is achieved when the angel is parallel to the wood direction, in our case it is 0°. This model is found to be an appropriate model for wood image identification enhancement and has a very good accuracy with an average value of 99:14%.
- Subjects
FEATURE extraction; GENETIC algorithms; ARTIFICIAL neural networks; RAIN forests; MANUFACTURING processes; TEAK; PROCESS optimization
- Publication
IAENG International Journal of Computer Science, 2019, Vol 46, Issue 2, p149
- ISSN
1819-656X
- Publication type
Article